document resume ce 002 692 walitz, louise berman · document resume. ce 002 692. walitz, louise...
TRANSCRIPT
ED 099 598
AUTHORTITLEINSTITUTION
SPONS AGENCY
REPORT NOPUB DATENOTE
AVAILABLE FROM
DOCUMENT RESUME
CE 002 692
Walitz, Louise BermanAn Analysis of the Labor Market for Policemen.California Univ., Berkeley. Inst. of IndustrialRelations.Manpower Administration (DOL) , Washington, D.C.Office of Research and Development.DLMA-91-05-70-13-1Sep 74298p.; Ph.D. Dissertation, California Univ.,BerkeleyNational Technical Information Service, Springfield,Virginia 22151
EDPS PRICE ME-$0.75 HC-$13.80 PLUS POSTAGEDESCRIPTORS Academic Achievement; City Problems; Economic
Research; *Employment Practices; *EmploymentQualifications; Equal Opportunities (Jobs); LaborEconomics; *Labor Market; Law Enforcement; *ManpowerNeeds; Minority Groups; Personnel Needs; *Police;Racial Discrimination; Recruitment; SexDiscrimination; Wages; Work Experience
IDENTIFIERS California; Oakland
ABSTRACTThe project examines shortages of police manpower and
analyzes the economic content of police hiring standards. A monopsonyexplanation of shortages is rejected in favor of a dynamic shortagemodel and a more rigorous model stressing the interaction of wages,hiring standards, authorized employment levels, and vacancy rates.Through mulitple regression techniques, the characteristics sought bythe Oakland (California) Police Department testing process between1965 and 1971 are defined. The probability of success is greatlyincreased as years of education increase, and is also raised byprevious police experience and high previous wages. It is decreasedby minority group status and local residence. Recent changes arereviewed, such as aggressive minority recruiting and affirmativeaction programs, which have greatly increased minority representationon the force. A review of the literature and a bibliography areincluded. (Author)
U S DI o'ANIMI NI Of TIT Al THEOUCAtioN n WI l f ARENATIONAL INSItTutE Or
I CHIC A ION
An Analysis of the Labor Market for Policemen
By
Louise Berman Wolitz
B.S. (Cornell University) 1963
DISSERTATION
Submitted in partial satisfaction of the requirements for the degree of
DOCTOR OF PHILOSOPHY
in
Economics
in the
GRADUATE DIVISION
of the
UNIVERSITY OF CALIFORNIA, BERKELEY
Committee in Charge
-- _...... .
i'.111..!RAPHIC DATASl-iEET
1. 1(;'',:t N".DIP.A
-,..)5-70-.1. 3 .1.
2.
Ij
3. 1%. i ili,t's A,.1..1:,:. N.,.
4. 1 air and Subtitle
An Analysis of the Labor Market. for Policemen
--4
5. P.cport DateSeptember 1071i
6
7. Autlimis)ki. Pcriorming Organiz.4tion Rept.
No.
9. Performing Organization Name and AddressUniversity of Californiainstitute of industrial 'elationsi:,erkeiey, California 94720
k10. Project Tas Worl: 1;nit NO.
il. Contract ,tlrara No.
DL91-05-70-13
1 2. ".p..nsor lug ilrptanizat ion Name And Addre%s
U.S. Department of Labor
Manpower AdministrationOffice of Research and Development1111 20th St., N.W., Washington, D. . 20210
13. Type oi Report & PeriodCovered
Final
14.
15. Supplementary Notes
16. Abstracts This project e:.amines shortages of police manpower and analyzes the economic
content, o police hiring standards. A monopsony explanation of shortages is
rejected in favor of a dynamic shortage model and a more rigorous model stressing
the interaction of wages, hiring standards, authorized employment levels, and
vacancy rates. Through multiple regression techniques, we attempt to define the
characteristics sought by the Oakland (California) Police Department's testing
process between 1965 and 1971. We find that the probability of success is
greatly increased as years of education increase and is also raised by previous
police experience and high previous wages. It is decreased by minority group
status and local residence. We review recent changes, such as aggressive minorityrecruiting and affirmative action prorrams, which have greatly increased minorityrepresentation on the force. A review of the literature and a bibliography are
included.
17. Key Words and Document Analysis. 17o. Descriptors
Aptitude tests Job description Professional personnel
Civil Servide Labor Quoiifications
Demand (economics) Local government Recruiting
Earnings Manpower Statistical analysis
Lc(,:.)mic analysis Manpower requirements Supply (economics)
Education (includes training) Negroes Tests
Ethnic groups Performance
Government employees. Performance evaluation
intelligence tests Personnel selection
17b. Identifiers/Open-Ended Terms
PolicePolicemen
17c. ( (),,A TI Fild/(group 5C , 51.
18.A,..allalllovtmement Distribution is unlimited.
Available from National Technical Information
Service, Springfield, Va. 22151.
19. Net urdy tlass (ThisReprtiLLA:,'41.1ip
21. No. ol Pages
i i-..
20. se. limy ( las. (1 hisPage
UNCLASSIFII'D
22. hit e
. ..- ,,,,,, rl I" tthal OilI MIA,/ kd 1.5.15 114E V. 3.711 TIIIS FORM MAY RFP1lODUCED
J J
BEST COPY IIVAILABLE
AN ANALYSIS OF THE LABOR MARKET FOR POLICEMEN
Louise Berman Wolitz
ABSTRACT
This project attempts to explain shortages of police man-
power and analyze the economic content of police hiring standards.
The shortages are of applicants considered qualified.
Chapter 2 outlines tho ambiguous and politically sensitive
nature of the polik_ job and the difficulties of defining succ:Jssful
performance.
Chapter 3 discusses three distinct theoretical models. The
simplest solution to shortages is to raise wages sufficiently to
eliminate them. Imperfect information and lags in institutional
decision-making are added to simple theory by the concept of
dynamic shortages, which allows wage adjustments to take con-
siderable, even infinite, time.
Secondly, shortages may arise from the exercise of monop-
sonistic market power, which implies an equilibrium position
with job vacancies at a monopsonistic wage.
Third is a rigorous model, stemming from a utility maxi-
mizing theory of public agency decision-making, which stresses
the complex interaction between wages, hiring standards,
11
authorized employment the value of the public service,
and expected vacancy rates. It is suggested that an agency may
wish to solve the problem of who shall be hired to fill vacant
positions, given authorized employment and wage levels, by
defining a minimum acceptable quality level chosen to maximize
the expected value of the net gain to society from the employees.
Chapter 4 analyzes the salaries, benefits, vacancies and
hiring standards of the Oakland, California, Police Department.
The evidence leads us to reject the monopsony explanation in
favor of a dynamic shortage model.
In Chapter 5, we attempt to define the characteristics
sought by Oakland's Mental Adaptability, Civil Service, and Oral
examinations, using data from 2,366 employment applications
submitted between 1965 and June, 1971, and multiple regression
techniques.
We find that while high school graduation is the only formal
educational requirement, the probability of success in the Mental
Adaptability is increased by large magnitudes as years of higher
education increase from some years of junior college through
three to four years of senior college or a college degree. Senior
college education still has a significant effect, moreover, on the
ability to pass the Civil Service test.
Local residence has a negative effect on the probability of
selection, perhaps because out-of-state candidates invest in more
search and self-selection.
Minority status, whether measured directly or indirectly
by the proxy of high school attendance in the South, has a large,
significant, negative effect on the probability of passing the
written tests, but a positive effect on passing the Oral.
Specific investment in police science courses is unimportant,
but previous police experience has a significant, positive effect
on the probability of success at all stages of the process.
The role of previous salary is clouded by the difficulties of
filtering out part-time from full-time work. The highest previous
salary category, $ 700 a month or more, has a significant positive
effect on the probability of success.
Age has a small, negative effect on passing the Mental
Adaptability, resulting from the well-known attrition of test-taking
ability with age, holding education constant, but a positive effect
on passing the Oral, reflecting concern with maturity.
Chapter 6 presents the dramatic changes in the legal en-
vironment, recruitment and selection between 1971 and 1974, re-
sulting, most importantly, in an increase in the minority per-
centage of the force from 3 percent in 1966, to 6. 5 percent in
iv
1970, to 16 percent in 1973, and 61 percent of the new hires in
1973. Affirmative Action and aggressive minority recruitment
have yielded significant results without sacrificing quality
standards. Moreover, patrolman and policewoman categories
have been merged, and height and weight standards abolished,
so more women should be hired. We also discuss Oakland's
turnover experience.
Chapter 7 concludes that wages, recruitment and selection
mus t be viewed together and hiring standards define not only the
potential police labor supply, but the nature of the job itself.
v
To my parents,
Abraham J. and Sylvia N. Berman
whose love and example provided the foundation
and to
my husband, Seth
and my daughter, Rebecca
who encouraged and supported the fulfillment
and to
the memory of my grandfather,
Israel Berman
vi
ACKNOWLEDGMENTS
This study would have been impossible without the continu-
ous encouragement, interest, generosity and good humor of
James M. Newman, Personnel Director of the City of Oakland.
He provided not only information and data, but also invaluable
insights. Extended thanks are due here as well to Charles Gain,
former Chief of Police of the City of Oakland, and George T.
Hart, the present Chief of Police, for their support of the study;
to Edward M. Toothman, of the State of California Commission
on Peace Officer Standards and Training, and a former City of
Oakland Chief of Police, for his early encouragement; and to
Lamar M. Vaughn and the staff of the Oakland Civil Service
Office for their assistance and their patience.
Professor Lloyd Ulman, chairman of my committee,
deepened my understanding of labor and manpower economics
and economic policy research, and supplied meaningful intellec-
tual support. His graduate seminars were excellent testing
ground for provisional research, and his comments were particu-
larly helpful. Professor Robert Aaron Gordon read the draft
carefully, asked excellent questions, and offered invaluable
guidance in combining economic rigor and economic policy.
vii
Professor William K. Muir shared his own deep understanding
of the police function, the Oakland Police Department, and the
relationship of intellectual theories to public policies, his
enthusiasm for the importance of the subject, and his ability to
clarify ideas. The interest and comments of the members of my
committee added greatly to the study. All remaining errors are
my own.
Professor Robert J. Flanagan deserves special thanks for
reading each chapter as it was completed and offering some
excellent suggestions for additional analysis. Other professors
and friends who have generously provided assistance, tangible
07 intangible, at various stages of this project include: Aaron.
Wildaysky and members of the Oakland Project, particularly
Frank Thompson and David Wentworth, Frank Levy, Stephen
Engleman, Michael Wiseman, George Strauss, Joseph Garbarino,
Sara A. Behman, Robert E. Hall, Frances Van Loo Flanagan,
Betty Heian, Geoffrey Car liner, Alan Fisher, David Lewin,
Thomas Thurston, and David Radin. I am greatly indebted to
Thomas Moore for expert programming and to Barbara Dewald
for truly extraordinary research assistance, and herculean
efforts in doing an enormous coding job with remarkable skill
and efficiency, aided by Martha Chase. Jean Georgakopoulos
viii
has been an excellent and dedicated typist. The superior, loving
child care provided by Bonnie Sussman and Rebecca Green was
an essential part of an intense summer of writing.
Joan Lewis and the staff of the Institute of Industrial Rela-
tions at the University of California, Berkeley, have provided
an office which is more like a gracious home. Gwendolyn Lloyd,
Clara Stern and Opal Pannell have offered indispensable library
service.
I am deeply appreciative of the financial support provided
by the Manpower Administration of the U. S. Department of
Labor which made this project possible. I would like to thank
the Computer Center at the University of California, Berkeley,
for the facilities to do the empirical work in this dissertation.
Without the consistent love, support and encouragement of
my husband, Seth, this effort could not have been completed.
My parents and my grandmother, Sophie Nelson, deserve credit
for instilling the belief, very early, that a woman must make a
place for herself in the world beyond the family.
ix
This report was prepared for the Manpower Administration,
U. S. Department of Labor, under research and development
grant No. 91-05-70-13. Since grantees conducting z.esearch and
development projects under Government sponsorship are en-
couraged to express their own judgment freely, this report does
not necessarily represent the official opinion or policy of the
Department of Labor. The grantee is solely responsible for the
contents of this report.
Copyright © 1974
by
Louise Berman Wolitz
x
TABLE OF CONTENTS
Chapter 1. INTRODUCTION 1
Chapter 2. THE JOB 15
DEFINITION OF THE JOB 15
POLICE PROFESSIONALISM 20
IDENTIFYING THE SUCCESSFUL POLICEMAN
EVIDENCE FOR OAKLAND 42
Chapter 3. THE THEORETICAL FRAMEWORK 59
THE SUPPLY OF POTENTIAL POLICEMEN 59
Calculation of Present Value of ExpectedLifetime Earnings 59
Age-Earnings Profile 61
Shape of Supply Curve 63
Uncertainty 63
Tastes 64
Noncompeting Groups 66
THE SUPPLY TO THE POLICE EMPLOYER 67
THE DEVINE MODEL 71
THE PROBLEM OF PUBLIC SECTOR DECISION-MAKING 77
THE EHRENBERG ANALYSIS 79
THEORY AND REALITY 90
xi
EMPIRICAL ISSUES AND ANSWERS 96
THE RELEVANCE OF THE THEORY 101
Chapter 4. THE INSTITUTIONAL FRAMEWORK:SALARIES, VACANCIES, HIRING STANDARDS. . 109
THE COMPOSITION OF THE OAKLAND POLICEDEPARTMENT 109
SALARIES 111
BENEFITS 129
VACANCIES 132
FACT AND THEORY 134
HIRING STANDARDS 137
Age 139
141
142
143
145
148
150
151
152
153
154
155
155
Height and Weight
V is ion
Education
Filing the Employment Application
Mental Adaptability Test
Civil Service Test
Physical Agility Test
Oral Examination
Medical Test
Psychiatric Evaluation
Background Investigation
Barriers to Entry
xii
Chapter 5. THE EMPIRICAL ANALYSIS OF SELECTIONTESTS 160
THE DATA 160
CENTRAL QUESTIONS 161
DEPENDENT VARIABLES 162
INDEPENDENT VARIABLES 162
Years of Education 162
Residence 164
Minority Status 166
Specific Police Training or Experienc e 167
Previous Wage 169
Number of Jobs Enumerated 170
Age 170
Physical Qualities 170
METHODOLOGY 171
RESULTS 176
Mental Adaptability 176
Civil Service Test 184
Oral Exam 185
Physical Agility 188
1971 Equations, Including a Minority Status Variable. .189
Chapter 6. THE CHALLENGES OF THE 1970s 196
MINORITY REPRESENTATION 196
The Legal Environment 196
Minority Composition of the Oakland PoliceDepartment 203
Administrative Attitudes 206
Research Efforts 208
A "Content Valid" Written Exam 211
Other Changes in the Selection Procedure 215
Active Recruitment of Minority Candidates 217
WOMEN 224
LESSONS FROM TURNOVER EXPERIENCE 229
Chapter 7. SUMMARY AND CONCLUSIONS 252
BIBLIOGRAPHY 264
xiv
TABLE OF TABLES
Table 2-1. Cohen and Chaiken Profiles 36
Table 2-2. Levy Profiles 43
Table 2-3. Levy Profiles 44
Table 2-4. Distribution of Oakland police officers byeducational level, August, 1962, toAugust, 1971 45
Table 2-5. Distribution of Oakland police officers byresidence, August, 1962, to August, 1971 46
Table 2-6. Turnover of Oakland police officers,August, 1962, to August, 1971 46
Table 2-7. Turnover of Oakland police officers byresidence, August, 1962, to August, 1971 46
Table 2-8. Turnover of Oakland police officers byeducational level, August, 1962, toAugust, 1971 47
Table 2-9. Positins taken by college graduates wholeft r)akland Police Department, 1956 to 1971 . . . 48
Table 3-1. Job applicants to Oakland Police Department,1965 to 1971, by occupational category 93
Table 3-a. Job applicants to Oakland Police Department,1965 to 1971, by occupational category,for Job 1, by years 94
Table 4-1. Oakland Police Department members, byrank, October 30, 1973 109
Table 4-2. Oakland Police Department members, bynumber of years of service, October 30, 1973 . . . 111
Table 4-3. Oakland Police Department members, bynumber of years of service and rank,October 30, 1973 112
XV
Table .4-4. Monthly salary of Oakland Police Departmentmembers, by rank, July, 1963, to July, 1974 . . .114
Table 4-5. Differences in salary of members of theOakland Police Department, by rank,July, 1963, to July, 1974 115
Table 4-6. Annual rate of increase Oakland PoliceDepartment patrolman and San Francisco-Oakland manufacturing production workers,July, 1963, to July, 1973 117
Table 4-7. Ratio of Oakland Police Department patrolmanand sergeant salaries to earnings of SanFranci sco-Oakland manufacturing productionworkers, July, 1963, to July, 1973 118
Table 4-8. Oakland Police Department salaries forpatrolman, step 1, patrolman, step 3, andsergeant, relative to median earnings in otheroccupational groups, 1969, for San Francisco-Oakland SMSA, total, Negro and Spanislsurname 121
Table 4-9. Oakland Police Department patrolman startingsalary relative to mean income of all SanFrancisco-Oakland males, 18 and older, andZvi to 34, in 1969, by years of school completedand race 127
Table 4-10. Oakland Police Department patrolman, step 3,and sergeant, salaries relative to mean incomeof all San Francisco-Oakland males, 18 andolder, in 1969, by years of school completedand race 128
Table 4-11. Total remuneration of Oakland patrolman,step 3, fiscal year July, 1970, to June, 1971. . . 130
Table 4-U. Oakland police officer vacancies, 1960 to 1973. . . 134
Table 4-13. Percent of all applicants appointed to theOakland Police Department, 1960 to 1970 137
xvi
Table 4-14. Common age requirements of California lawenforcement agencies in 1968 140
Table 4-15. Common height and weight requirements ofCalifornia law enforcement agencies in 1968. . . . 142
Table 4-16. Common visual acuity tandards of Californialaw enforcement agencies in 1968 143
Table 4-17. Number of employment applications for Oaklandpatrolmen filed and accepted, 1960 to 1970 147
Table 4-18. Number of Oakland patrolman applicants takingand passing the Mental Adaptability Test,1960 to 1970 150
Table 5-1. Percent of applicants in each non-referencecategory successful on the Mental Adaptability . . 177
Table 5-2. Regression coefficients, all applicants, 1965to June,1971 178
Table 5-3. Percent of applicants in each categorysuccessful on the oral exam 186
Table 5-4. Percent of 1971 applicants in each non-referencecategory successful on the Mental Adaptability . . 190
Table 5-5. Percent of white and minority members ineach non-reference category 191
Table 5-6. Regression coefficients: January to June,1971, applicants 193
Table 6 -1. Percent of new appointees from minority groups,Oakland Police Department, 1961 to 1973 204
Table 6-2. Total number and percent of Oakland PoliceDepartment positions filled by each minoritygroup, December, 1966, to December, 1973. . . 205
Table 6-3. Oakland Police Department members, by dateof appointment and ethnicity, October 30, 1973 . . 207
xvii
Table 6-4. Experience of ethnic groups on patrolmanexamination, Janu;,ry, 1972, to September,1973 214
Table 6-5. Educational levet of individuals who resignedfrom the Oakland Police Department, July,1966, to June, 1971 232
Table 6-6. Type of job taken by individuals who resignedfrom the Oakland Police Department, July,1966, to June, 1971, by educational level 233
1
CHAPTER 1
INTRODUCTION
On July 23, 1965, "recognizing the urgency of the Nation's
crime problem and the depth of ignorance about it,"1 President
Lyndon. Johnson established the Commission on Law Enforcement
and Administration of Justice, through Executive Order 11236.
In its omnibus 1967. report, The Challenge of Crime in a Free
Society, the Commission emphasizes unequivocally the central
role of police manpower policies in the functioning of the Ameri-
can criminal justice system. "The police personnel need that
the Commission has found to be almost. universal is improved
quality ... The recommendations that have been made... about
community relations and policymaking, and the ones about...
organization and management, are predicated on the sharp
improvement of the quality of police personnel from top to
bottom." 2 In its much-quoted Table of Recommendations, the
Commission continues:
Widespread improvement in the strength and caliber
of police manpower, supported by a radical revision
of personnel practices, are the basic essentials for
achieving more effective and fairer law enforcement,
2
Educational requirements should be raised to college
levels and training programs improved. Recruitment
and promotion should be modernized to reflect educa-
tion, personality and assessment of performance. The
t raditional monolithic personnel structure must be
broken up into three entry levels of varying responsi-
bility and with different personnel requirements, and
lateral entry into advanced positions encouraged.3
Other discussions of police manpower abound with widespread
complaints of "shortages" of personnel to fill authorized positions 4
and difficulties in recruiting and selecting desirable candidates. 5
These problems are enhanced by serious controversies over
appropriate hiring standards and racial/ethnic composition and
their interrelationship.
While the intentions of the Commission are laudatory, the
clarity and sureness of its statement is deceptive. To speak of
"improvement" implies a perception of an existing state and of a
better one towards which one seeks to move. Yet, as recognized
by the Commission itself in its Foreward, discussions are being
held and policy recommendations are being made in an abysmal
analytic vacuum, not just about ideals but also about realities.
Are there, in fact, "shortages' ,6 of police personnel? If so,
3
why, and how an they be corrected? If not, why arc- they
widely perceived by police departments? A classic economic
discussion would infer that if there are shortages, perhaps
salaries are too low. Would raising police salaries be a good
policy approach? How can persistent complaints of "shortages"
be reconciled with the equally persistent debates about the
existence of arbitrary hiring standards, discrimination against
minorities, and demands to upgrade the "quality" of the men
hired, all phenomena more usually associated with freedom to
ration scarce jobs among many potential workers? Perhaps a
partial explanation lies in the very murky, institutional area of
labor market adjustment where hiring standards, recruitment
and selection procedures are worked out and administered.
Even more puzzling is the question of what is meant by
improving the "strength," "caliber," or "quality" of police man-
power. How do we know that raising educational requirements
to college levels will improve police quality? What personality
traits are we seeking in our policemen? How do we assess
police performance? How do we design recruitment, selection
and promotion processes to yield "quality" men?
It is the purpose of this study to shed some substantive
light on these questions. Immediately, it became apparent that
4
"shortages, if any, were caused not by too few applicants
offering their services, but rather by an insufficient number
deemed "qualified" by police departments.
Although police departments are encountering difficulty
in hiring new officers, this is not due to a dearth of
applicants, but to a lack of qualified applicants. In
1961, only 22.3 percent of the applicants for positions
in 3b8 police departments were accepted. The applicant
success rate in many departments is far lower. For
example, in 1965, only 2. 8 percent of the candidates
for the Los Angeles Police Department were eventually
accepted into the force. In 1966, only 29 of 1,033
applicants were hired by the Dallas Police Department.
A survey of police vacancies in California clearly "indicated
generally that most agencies had no difficulty in attracting men
in quantities, the problem was hiring and retaining men of
quality. "8
One is forced, therefore, to find out who it is that police
departments actually hire and why. What qualities do they seek
from their labor supply and why? In an environment in which st,
little is known, this question is complementary to the more
normative one of what qualities should they be seeking and why.
7
5
A meaningful analysis of the police labor market must focus on
a conception of the economic adjustment process which is particu-
larly sensitive to the role of such non-wage factors as recruit-
ment, selection and training.
To understand the role of non-wage economic factors, it is
necessary to study particular departments, their decisions,
decision rules and institutions. We were fortunate enough to
obtain the interest and cooperation of the Police Department and
Civil Service Commission of the City of Oakland, California.
The empirical focus of this study will be an attempt to assess the
role of the recruitment and selection process in yielding a labor
supply to the City of Oakland to fill its demand for patrolmen.
In particular, we will seek to discover the determinants
of success for Oakland Police Department applicants at three
critical stages of the process: the Mental Adaptability, Civil
Serice and Oral exams. We will concentrate on those variables
that labor market theories lead us to believe should be important,
leaving the more difficult, perhaps even more important, area
of personality assessment to the psychologists. A priori, we
expected that the critical variables would be: years of education,
residence, racial/ethnic status, specific police training or
experience, previous wages earned, employment history and age.
6
We test, through multiple regression techniques, the influence
of these variables on 1) the probability of success and 2) the
score on each of the exams. We will summarize our results
here, briefly and only suggestively.
We find that while high school graduation or its equivalent
is the only formal educational requirement for applicants to the
Oakland Police Department, the probability of success at the
first stage of the testing process, the Mental Adaptability, is
increased by large magnitudes as years of higher education
increase from some years of junior college through three to four
years of senior college or a college degree. Senior college ed-
ucation still has a significant effect, moreover, on the ability to
pass the Civil Service test. One might argue, therefore, that
the Oakland Police Department should assess carefully the
advantages and disadvantages of making explicit their implicit
educational standards by increasing the minimum educational
requirement. They would thus transmit more realistic signals
of their expectations to potential applicz:if'.,. They would
acknowledge more clearl y their competitors for labor. The
wage necessary to attract an adequate supply of college graduates,
for example, would differ considerably from that necessary to
attract the same number of high school graduates. Perhaps most
importantly, they would save the substantial cost of recruiting,
screening and testing applicants very likely to be rejected.
Alternatively, one can conclude that perhaps the emphasis
placed by tests upon years of education is screening out desirable
candidates who would make "successful" policemen, and it is the
tests, not the standards, which should be re-evaluated. These
and other issues will be discussed mo2,! fully later.
Local residence appears on balance to have a negative
effect on the probability of selection. This may be partly ex-
plained by the fact that it is considerably more costly for an out-
of-state resident to seek out the opportunity and to apply than it
is for a local fellow to walk into City Hall. off the streets. There-
fore, out-of-state people must have, on the average, more motiva-
tion or interest, and would do more self-selection than local
residents. This may argue either in favor of increased national
recruiting efforts to improve the average "quality" of applicants
as perceived by the department, or against the perception of
quality employed.
Minority group status, whether measured directly or
indirectly by the proxy of high school attendance in the South,
has a large and significant negative effect on the probability of
passing the standardized written test. It should be emphasized
8
here that multiple regression techniques allow us to estimate the
independent effect of minority status, holding everything else
(education, residence, etc.) constant. This points either to
intentional discrimination or to the validity of the complaints of
the minority community that such tests are culturally biased and
thus inherently discriminatory. In fact, recent court decisions,
including one by the Supreme Court of the United States, in
related cases, make it clear that if a selection test is inherently
discriminatory, the burden of proof is on the testing agency to
show that the test predicts performance. If predictive validity
cannot be demonstrated, the test must be abandoned and, some-
times, compensatory measures taken to correct its effects. 9
it should be added here that Oakland has recognized the
desirability and/or need to increase minority representation.
This is demonstrated by the fact that, at least in 1971, the last
year of our study, those minority members who make it through
the written tests have a higher probability of passing the oral
than non-minority candidates. More recently, Oakland has
switched from national recruiting to residence preference,
adopted an affirmative action program and embarked on aggressive
minority recruitment campaigns which have yielded significant
results.
9
Specific investment by applicants in college or non-college
police science courses is unimportant. Neither the Police
Department nor the Civil Service staff believe outside training to
be worth anything. They emphasize their own intensive 20 week
recruit school. Prior experience as a policeman, however, does
have a significant positive effect on the probability of success at
all stages of the process. Of course, such candidates have already
succeeded in a similar selection process elsewhere.
The role of previous salary as an indicator to the depart-
ment of an applicant's potential performance as a policeman is
clouded in our study by difficulties in filtering out part-time from
full-time work. The highest previous salary group, $700 per
month or more, does have a consistent significant positive effect
on the probability of a man being accepted. Other salary groups
were inconclusive. To minimize the distorting effects of part-
time work, we looked at the non-college group alone. For this
group, the previous wage variable behaved as economic theory
would predict, with higher salaries indicating progressively
greater probabilities of being selected. Previous salary does
indicate alternative worth and promise as a policeman. Number
of previous jobs enumerated on the application is unimportant.
Age has a small negative effect on the probability of passing
10 .
the Mental Ada Dta.bility, resulting from the well-known attrition
of test-taking ability with age, holding education constant. Age
has, however, a positive effect on success on the oral, reflecting
a concern with maturity.
We propose to proceed as follows. Chapter 2, "The Job,"
will describe the' unique nature of the job of policeman, the diffi-
culties in defining a "successful" policeman, the current contro-
versies, and the importance of the definition of the job to the
operation of the labor market.
Chapter 3, "The Theoretical Framework," will seek a
theoretical explanation of the alleged shortages of police man-
power in the 1960's. It will review the determinants of the supply
of labor to police departments. It will describe in detail three
distinct theoretical models which might be helpful in understand-
ing Oakland's experience.
First, the simplest economic reasoning argues that short-
ages should be eliminated by a sufficient rise in relative wages.
Informational imperfections and institutional decision-making
are added to the simple model by the conception of dynamic
shortages, which allows wage adjustments to take considerable,
even infinite, amounts of time.
11
Second is a model which relies on market power, specif-
ically monopsony, to explain shortages as consistent with a
monopsonistic equilibrium position and a monopsonistic wage.
Police departments experience job vacancies in equilibrium
because of their exercise of moncpsony power in the labor
market.
Third is a rigorous model, stemming from a utility max-
imizing theory of public agency decision-making, which stresses
the complex interaction between wages, hiring standards,
authorized employment levels, and vacancy rates. The problem
faced, in fact, by real world police agencies is defined: given
authorized employment and wage levels, who shall be hired to
fill vacant positions? It is suggested that the agency may wish
to approach this problem by defining a minimum acceptable
quality level, perhaps through a vector of hiring standards,
chosen to maximize the expected value of the net gain to society.
The formal implications of this model are discussed in detail.
Chapter 4, "The Institutional Framework: Salaries,
Vacancies, Hiring Standards, H analyzes the salaries, vacancies
and hiring standards of the Oakland Police Department. The
evidence leads us to reject the monopsony explanation in favor
of a dynamic shortage model.
12
Wv. also begin our effort to define and describe
the vector of hiring standards of the Oakland Police
Department.
Chapter 5, "The Empirical Analysis," which we have
already discussed, further specifies the vector of hiring stand-
ards by analyzing the characteristics for which the Oakland
Police Department's selection process between 1965 and 1971 is
testing.
Chapter 6, "The Challenges of the 1970's," presents the
dramatic changes in recruitment and selection which have
occurred between 1971 and August, 1974, and their effect on the
composition of the Oakland police force. Residence requirements,
Affirmative Action, and aggressive minority recruitment cam-
paigns have yielded significant results. The minority composi-
tion of the force has increased from 3 percent in 1966 and 6. 5
percent in 1970 to 16 percent in 1973 and 61.0 percent of new
hires in 1973. The sexual composition of the police force is
changing as well, as patrolman and policewoman categories are
merged, and height and weight standards which would exclude the
majority of women from the new police officer job are abolished.
In Chapter 6, as well, we also discuss the turnover experience
of the Oakland Police Department and analyze in depth why
13
officers resignk We discuss and reject the possibility of
increasing economic incentives to retain resignees.
Chapter 7, "Summary and Conclusions, " concludes that
our evidence is consistent with the model of dynamic shortages
and inconsistent with the monopsony power explanation of job
vacancies in equilibrium. We do not v.,sh to rest with only the
simple model, however, because the complex interactions
between wages, hiring standards, authorized employment levels
and vacancy rates must be explored further. We hope we have
made some contribution to observing, analyzing and under-
standing the central role of the choice of the vector of hiring
standards.
14
CITA PTER 1
FOOTNOTES
1. President's Commission on Law Enforcement and Administra-tion of Justice, The Challenge of Crime in a Free Society(Washington, D. C. : U. S. Government Printing Office, 1967),Foreward.
2. Ibid. , p. 107.
3. Ibid. , pp. 29 4-295 .
4. See, for example, Raymond I . Bancroft, "Municipal LawEnforcement, 1966," Nation's Cities, February 1966, pp. 15-17; and President's Commission on Law Enforcement andAdministration of Justice, Task Force Report: The Police,pp. 133-134.
5. See, for example, State of California, Commission on PeaceOfficer Standards and Training, California Law EnforcementRecruitment Program, p. 4.
6. The question of defining "shortage" precisely will be dealtwith at length in Chapter 3.
7. President's Commission on I aw Enforcement and Administra-tion of Pistice, Task Force Report: The Police, p. 134.
8. State of California, Commission on Peace Officer Standardsand Training, California Law Enforcement RecruitmentProgram, p. 4.
9. The Supreme Court decision is Griggs v. Duke Po,er Com-pany, March 8, 1971. See also Julius Arrington et al. v.Massachusetts Bay Transportation Authority, Civ. A. No.69-681, United States District Court, Massachusetts, December22, 1969: Carter et al.. vs. Hugh Gallagher et al. , UnitedStates District Court, District of Minnesota, Fourth DivisiOn,March 9, 1971: United States of America v. Virginia Electricand Po,.ver Company et al. , Civ. A No. 638-70-R, UnitedStates District Court for Eastern District of Virginia,Richmond Division, May 4, 1971.
15
CHAPTER 2
THE JOB
DEFINITION OF THE JOB
Unlike most jobs in our increasingly specialized economy,
the job of policeman remains inordinately difficult to define.
The policeman is simultaneously technician and generp.list. In
the routine performances of his daily tasks, he must adhet. to
finely drawn rules of dress, report writing, obedience to command,
weapons management, citizen contact procedure and law enforce-
ment procedure. He. must often endure long hours of boredom,
watching, waiting, appearing in court, filling out endless forms.
Yet, inherent ia the job is the knowledge that at any moment,
virtually ,Tkithout warning, he may be called upon to exercise
unlimited disc retio spontaneously and without guidance, in
matters of life and death for others and himself. He must be an
efficient cler,, without the security of constant routine, a skilled
craftsman without clear quality standards or freedom to improvise,
a professional without certification, systematic research or
collegial self-government.
The complexity of the police task was well captured in 1931
by the Wickersham Commission.1 We have defined it no better
16
today.
Reviewing the tasks we expect of our law enforcement
officers, it is my impression that their complexity is
perhaps greater than that of any other profession. On
the one hand we expect our law enforcement officer to
possess the nurturing, caretaking, sympathetic,
empathizing, gentle characteristics of physician, nurse,
teacher, and social worker as he deals with school
traffic, acute illness and injury, juvenile delinquency,
suicidal threats and gestures, and missing persons.
On the other hand we expect him to command respect,
demonstrate courage, control hostile impulses, and
meet great physical hazards. . . He is to control
crowds, prevent riots, apprehend criminals, and chase
after speeding vehicles. I can think of no other pro-
fession which constantly demands such seemingly
opposite characteristic s.
Research interest in the policeman's job was galvanized by
the explosive urban and campus tensions of middle-1960's America
and the questions raised by the very visible and public police
response. The urgency of remedying our abysmal ignorance of
policeman was apparent. The President's Commission on Law
17
Enforcement and Administration of Justice was therefore estab-
lished by President Johnson on July 23, 1965, to investigate
every facet of crime and law enforcement in America." 2 It
issued its omnibus report, The Challenge of Crime in a Free
Society, in February, 1967, followed by detailed Task Force
reports on each phase of the criminal justice system and also
staff research and consultants' studies. The literature, both
descriptive and analytic, has blosso:Tied since then, and if our
questions have not been answered, at least many have been
raised. The difficulty of merely defining and describing the
policeman's job is underscored by the appearance of so many
recent books by major publishers with this one aim in mind.
I am emphasizing this issue here because a labor market study
would logically begin with a clear and concise definition of the
job to be studied. The fact that the job defies simple definition
is the single most important characteristic of the labor market
and sets the stage for many of the controversies and problems
upon which this paper focuses. The reader interested in a fuller
discussion of the details, conflicts and problems inherent in the
policeman's job is directed to the following books, listed in
approximate order of appearance: David J. Bou-lua, The Police
(New York: Wiley, 1967); Arthur Niederhoffer, Behind the Shield
18
(Garden City: Doubleday, 1967); Jerome Skolnick, Justice
Without Trial (New York: Wiley, 1967): James Q. Wilson,
Varieties of Police Behavior (Cambridge: Harvard University
Press, 1968); George E. Berkeley, The Democratic Policeman
(Boston: Beacon Press, 1969); Paul Chevigny, Police Power
(New York: Vintage Books, 1969); Gene Radano, Walking the
Beat (Toronto: Collier Books, 1969); L. H. Whittemore, Cop.
(New York: Holt, Rinehart and Winston, 1969): William A.
West ley, Violence and the Police (Cambridge: MIT Press,
1970); Albert J. Reiss, Jr. The Police acid the Public (New
Haven: Yale University Press, 1971); and Jonathan Rubinstein,
City Police (New York: Farrar, Straus and Giroux, 1973).
The functions of the police are officially described as 1)
protection of life and property, 2) preservation of the peace,
3) prevention of crime, 4) detection and arrest of law violators,
5) enforcement of laws and ordinances , and 6) safeguarding the
rights of individuals. 3 Ambiguity and potential conflicts are
already obvious. Law enforcement functions may clash with
order maintenance ones. Prevention and detection of crime may
clash with individual rights. Whose conception of peace or order
or conformity is to be preserved? Should the law be bent to
preserve order, or order reign over law? Since each case a
19
r,:inciles is different, there can be no exhaustive
catalogue of how he should proceed to carry out his mandate.
At best, he is told what behavior to avoid. Substantial individual
discretion and selective enforcement of the law are inevitable.
And in his person coexist two potentially combustible elements,
danger and authority, exacerbated by a constant pressure to
appear efficient. In addition, the primary duties are sometimes
forgotten under the pressure of accomplishing various licensing,
inspecting and administrative tasks.
There are several research studies in progress now to find
out exactly what it ,is that policemen do with their time. Prelim-
inary evidence from these studies seems to bear out that in spite
of the drama of the above description, policemen actually spend
only approximately ten percent of their time on law enforcement
related activities. The bulk of their time is spent on service
related activities, followed by order maintenance and information
gathering. 5
The most serious question which arises from ambiguity is
that of responsibility. To whom is a policeman, a police chief,
a police force ultimately responsible? The locality it serves
provides it with no exhaustive daily guidance. Local governing
boards, in fact, become involved only in clear cases of police
20
failure. Most localities today, particularly urban ones, are not
homogeneous entities, but represent diverse political and social
groups with widely disparate, often passionately held views as
to the appropriate role of the police. Whose views prevail') How
are they determined? Can a majority view be defined? Is it
middle class, Catholic, intellectual, young, middle-aged,
Democratic, Republican, white, black, etc.? Are all groups in
the community involved in the determination, or only some?
What about minority rights?
These are rhetorical questions, but critical ones. For
some cities there are answers, for others there are not. Many
cities simply attempt to avoid the questions. For our purposes,
it must be clear that the politics of the police, and its social
mandate, are heavily controversial, fundamentally explosive
subjects, totally unresolved in America today. Many of the
issues of police manpower we will be dealing with in economic
terms here are really reflections of or attempts to mask a
powerful political debate. So we must remain alert for implica-
tions broader tliar. our subject.
POLICE PROFESSIONALISM
One example, extremely critical to this study, is the issue
of professionalism or professionalization of police forces. Indeed,
L1
one way to divert the political questions is to make the police a
profession, like doctors, lawyers, teachers, responsible pri-
marily to themselves, their collegially determined goals and
standards of conduct. Professionalism itself is difficult to
characterize. 6 In Durkheim's view, what is distinctive about
professional groups is not merely their high status, high skill,
politically important monopoly over certain kinds of work, or
particular structure of control over work, but, most importantly,
the infusion of work and collective organization with moral values
and the use of sanctions to insure that these moral values are
upheld. An alternative concept of professionalism is associated
with a managerial view emphasizing rationality, efficiency and
universalism. This view envisages the professional as a bureau-
crat, almost as a machine calculating alternative courses of
action by a stated program of rules and possessing the technical
ability to carry out decisions irrespective of personal feelings.
This model has particular disadvantages when applied to the
police because it fails to bridge the gap between the maintenance
of order and the rule of law and because the policeman is an
especially nonmechanical official.7
The argument that policemen should become professionals
rests on inc reasing awareness of the complexity of the job, the
responsibility and judgment involved, the critical and isolated
nature of some of the decisions required. Education is proposed
as the vehicle through which men can be molded to handle dis-
cretion and make good decisions. The quality of policemen
should therefore be upgraded, two-year or four-year college
degrees required, departmental training programs deepened and
broadened, career hierarchies established, salaries raised,
pseudo-military conformity revised, physical prowess de-
emphasized, selection implements refined. All of this, more-
over, should improve the image, status and power of the police-
man within his community and secure public confidence and
support. Although the rhetoric encompasses both definitions of
professionalism, the reality tends towards the strictly adminis-
trative one.
In thoughtful, improvement-oriented police circles,
increasing professionalism is often proposed as a rational cure
for the problems of relations with minority communities, cor-
ruption, securing quality personnel and assuring public confidence,
particularly in the context of ever more sophisticated technolog-
ical advances and legal questions. The reports of the President's
Commission on Law Enforcement and Administration of Justice
unanimously and enthusiastically support and make extensive
23
policy recommendations in this direction, as quoted on the first
page of the Introduction to this study. The suggestions most
often offered can be summarized as follows:8
1) high standards of admission, through mandatory
minimum educational, mental, moral and physical
requirements;
2) extensive training program) for candidates, perhaps
with mandatory minimum standards, uniform curricula,
and specially trained teachers;
3) a career hierarchy, with progression determined by
education, training and experience levels;
4) licens;ng of members;
5) a code of ethics;
6) a special body of knowledge and theory, improved by
organized research efforts;
7) altruism and dedication to the service ideal;
8) autonomous control;
9) pride of the members in their profession;
10) publicly recognized status and prestige;
1) coordinated efforts of local jurisdictions to meet
common administrative, recruitment and selection
problems.
L4
California police departments, arid the Oakland Police Depart-
ment in particular, are proudly pointed out as having already made
substantial progress towards becoming professional departments.
Within police departments themselves, the issue of pro-
fessionalization is still controversial. The street-wise, old-time,
experienced beat patrolman who could not meet the suggested
educational standards generally scorns and resents both the
standards and their "cocky," inexperienced human embodiments.
Considerable hostility and social tension may be generated
between the young, college-educated, professional and the old-
timer.
The doctrine of the professional movement advocates
education, a non-punitive orientation, strict legality
in every phase of police work, and, most important
of all, good public relations. These principles create
an atmosphere in which only the professionally
directed police officer can operate successfully. On
a lower plane within the hierarchy exists the tough-
minded, authoritarian working class of the police
society. Conditions here foster exaltation of power
and force and contemptuousness of the unrealistic,9soft-minded, subversive ideas of professionals.
25
Outside scholars of the police such as James 0. Wilson,
Arthur Niederhoffer, and Jerome Skolnick are also not convinced
by the professionalization rhetoric, agreeing that although police
often have the responsibility, discretionary power, and aspira-
tions of professionals, they are not professionals yet.
Occupations whose members exercise, as do the
police, wide discretion alone and with respect to
matters of the greatest importance are typically
"professions"--the medical profession, for example.
The right to handle emergency situations, to be privy
to 'guilty information', and to make decisions involving
questions of life and death or honor and dishonor is
usually, as with a doctor or priest, conferred by an
organized profession. The profession certifies that
the member has acquired by education certain informa-
Con and by apprenticeship certain arts and skills that
render him competent to perform these functions and
that he is willing to subject himself to the code of
ethics and sense of duty of his colleagues (or, in the
case of the priest, to the laws and punishments of
God), Failure to perform his duties properly will,
if detected, be dealt with by professional sanctions--
26
primarily, loss of respect. Members of professions
tend to govern themselves through collegial bodies,
to restrict the authority of their nominal superiors,
to take seriously their reputation among fellow
professionals, and to encourage some of their kind to
devote themselves to adding systematically to the
knowledge of the profession through writing and
research. The police are not in any of these senses
professionals. They acquire most of their knowledge
and skill on the job, not in separate academies; they
are emphatically subject to the authority of their
superiors; they have no serious professional society,
only a union-like bargaining agent; and they do not
produce, in systematic written form, new knowledge
about their craft.
In sum, the order-maintenance function of the
patrolman defines his role and that role, which is
unlike that of any other occupation, can be described
as one in which SUB-PROFESSIONALS, WORKING
ALONE, EXERCISE WIDE DISCRETION IN MATTERS
OF UTMOST IMPORTANCE (LIFE AND DEATH, HONOR
AND DISHONOR) IN AN ENVIRONMENT THAT IS
27
A PPR EHENSIVE AND PEHI IA PS HOSTILE. The
agents of various other governmental organizations
may display one or two of these characteristics but
none or almost none display all in combination.
The doctor has wide discretion over matters of life
and death, but he is a professional working in a
supportive environment. The teacher works alone
and has considerable discretion, but he may be a
professional and in any case education, though
important, is not a matter of life or death. A welfare
worker, though working alone among apprehensive
clients, has relatively little discretion--the laws
define rather precisely what payments he can
authorize to a client and supervisors review his
written reports and proposed family budgets.10
The issue of the possibility and desirability of police
professionalization could be passed over as of purely theoretical
interest if attempting to apply the concept had few costs. How,
after all, could ungrading the quality of police forces be a bad
policy? Clearly, there is an economic cost. Briefly, higher
salaries must be paid in order to attract a more highly "qualified"
labor force. A college educated force must co ( more than a
28
high school educated one. If quality standards are artificially
high, so are salaries. If departments do not succeed in attracting
enough men of "professional quality," they will be operating
chronically below their authorized strength, a prevalent situation
in the 1960's.
Moreover, additional hiring standards increase barriers
to entry to the job, reducing the effective supply of labor. This
is critically important here because increasing educational stand-
ards is particularly likely to reduce the already grossly inadequate
pool of applicants from minority groups. Professionalization
directly clashes with, even makes a mockery out of, loudly
espoused efforts to increase the representation of minority groups;
yet minority status itself may determine professional efficiency
The serious under-representation of minority groups on
police forces is well documented in every city, county and state
where statistics are available. 11 For example, according to a
suit filed by the United States Justice Department on August 15,
1973, in U. S. District Court in Chicago, the Chicao Police
Department has about 13,500 police officers, of which 16 percent
are black and 1 percent are Spanish-surnamed. About 33 per-
cent of Chicago's population is black and 7 pc rcent Spanish
surnamed. Un the same day, the Justice Department filed a suit
1
29
in Federal court in Buffalo, which employs 1, 405 uniformed
police officers. Of the uniformed officers, 1. 8 percent are
black, none is Puerto Rican. Blacks make up more than 20 per-
cent of Buffalo's population and Puerto Ricans 0. 4 percent.12
In San Francisco, 9 percent of the city's police force is composed
of minority members. The 1970 Census listed minorities as 43
13percent of the population. Washington, D. C. has one of the
highest proportions of black officers in the country, 18. 5 percent,
with a population 54 percent black.14 In Oakland, which is almost
50 percent minority, 7. 3 percent of the uniformed police force was
minority as of June 1970. 15 Oakland's record is particularly bad
in a generally bad picture. to
Advocates of police professionalization, including the
President's Commission, have not dealt openly with the direct
conflict between raising educational standards and improving
minority representation. Some even advocate that if police
forces become professional, the problem of minority representa-
tion becomes unimportant because a professional force, by defini-
tion, would enforce all interests equally, making representation
of "special" interests unnecessary. The political naivety, at
best, of this view requires no elaboration.
In fact, the opposite case can be made persuasively. The
30
current demand for a more formally educated police force may
be in part a drive by the growing cl.:ss of professional and
educated Americans to gain influence over police actions, to
make "them" more like "us, " to insure responsiveness to the
perceived needs of this class. 17 The whole conflict over police
aims and standards can be looked at politically as a theater in
which different population groups compete for representation
and influence. The issues, perhaps, cannot be handled until
police departments and political leaders explicitly recognize and
deal with, rather than implicitly obscure, the political context
of police work.
These profound basic political controversies must be kept
in mind as one turns to the economic and empirical issues. The
controversies are all the more serious because the professional
standards proposed have no clear empirical validity in identify-
ing successful policemen.
IDENTIFYING THE SUCCESSFUL POLICEMAN
The empirical social scientist offers another means of
avoiding political controversy. The qualities needed to be a
successful policeman can be determined empirically with the
aid of modern statistical techniques, and selection procedures
can then be designed to choose those applicants most likely to .
31
succeed. A requirement that all police candidates have a college
degree, for example, can be recommended, validated and en-
forced if it can be demonstrated statistically that college grad-
uates perform better on the job. Indeed, advocates of police
professionalization often leave the impression that such valida-
tion has already taken place.
A burgeoning interdisciplinary literature has been emerg-
ing addressed to this effort. It varies in quality from the
bizarre to the careless to the sophisticated and careful. Bernard
Cohen's and Jan M. Chaiken's Fblice Background Characteristics
and Performance, published by the New York City Rand Institute
in August 1972, represents by far the best study. We will con-
sider it here in detail as illustrative of the potential role of
competent, thoughtful empirical analysis in identifying desirable
characteristics of policemen.
Cohen and Chaiken obtained and analyzed background and
performance data for 1, 915 officers appointed to the New York
City Police Department in 1957, 1,608 of whom were still
active members in 1968 when the information was collected.
From 18 possible performance measures they identified by
factor analysis five performance patterns:
1) termination, which describes the officer who left
32
the Department prior to 1968, either voluntarily
or involuntarily;
2) career advancement, which refers to the officer
who obtained special assignments or promotions,
frequently coupled with above-average numbers
of awards;
3) departmental discipline problem, which describes
the officer who had an above-average number of
departmental charges, and frequently also had
an above-average number of times sick:
4) above-average number of civilian complaints; and
5) above-average number of allegations of harassment.
The last four performance patterns were found to be inde.)endent
of each other.
They then explored the relationship between these per-
formance patterns and the following background characteristics
of the officers:
A) Early background characteristics
1) Race
2) Age
3) I. Q.
4) Civil Service Test Score
18
33
5) Region of birth
6) Siblings
7) Occupational history
8) Military history
9) Arrest history
10) Civil court appearances
11) Large number of debts
12) Prior history of psychological disorders
13) History of mental disorder in family
14) Father's occupation
15) Number of residences or place of residence
16) Marital status and number of children
17) NumbeA of summonses
18) Background investigator's rating
B) Later background characteristics
1) Recruit training score
2) Probationary evaluation
13) Education 9
Through multiple regression analysis, they estimated the
average performance levels for officers having specified combi-
nations of background characteristics and identified the charac-
teristics which made the greatest contribution to explaining
34
variations In performance among officers. For white officers,
the strongest explanatory variables, in order, were: average
grade in police academy training courses, probationary evalua-
tion, background investigator' s rating, military disciplinary
record, employment disciplinary record, level of education,
number of appearances in civil court, age at time of application,
civil service examination score, marksmanship, prior arrest
history, and 1. Q. For black officers, the strongest two pre-
.dictors were recruit training score and probationary evaluation. 20
Background characteristics 11 to 17, for example, were found to
be unrelated to later performance, in spite of an a priori
expectation that they would be negatively related. The policy
implication is, therefore, that these factors should not be con-
sidered in the selection process.
To determine the extent to which performance measures
would he expected to vary with background characteristics,
they calculated some typical values of performance measures
from regression equations. For example, they found that a
hypothetical candidate with three military or employment
disciplinary incidents, the lowest possible recruit score, and
two "unsatisfac,my" marks on his probation report would be
expected to have 8. 5 times as many substantiated complaints of
35
misconduct as a man with no military or employment discipline
record, a recruit score of 90, and no "unsatisfactory" marks on
probation. A similar disparity in civilian complaints was found
between older college graduates and younger high school grad-
uates. 21
Using the results from cross-tabulations and regression
analysis, they developed profiles of the candidates who are most
likely to embody the performance characteristics identified in
the factor analysis. The profiles differed for white and black
officers as shown in Table 2-1.
Three hundred and seventy-six out of 2, 002 men appointed
in 1957 (or 19 percent) had left the department by 1968. Voluntary
and involuntary terminations could not be distinguished from the
data. The peaks of the terminations tended to occur at two years
(18 percent) and five years (12 percent) after appointment. The
men who left the force were younger, if married had fewer
children and debts, and were considerably better educated than
those who remained.
It is particularly noteworthy that one-third of the
college-educated recruits in 1957 (8 out of 24) were
found to have left the force by 1968, compared to
19 percent of the men who had not graduated from
36
Table Cohs.n and Chaiken )files 22
1. The men most likely to be a discipline problem for thedepartment, with a large number of departmental chargesand times sick, had the following characteristics:
Whites BlacksYoung at appointment High I. Q.Non-college graduate Few siblingsExcessive summonses and Poor background rating
debts Low recruit scoreEmployment disciplinary Poor probationary evaluation
record Born in New York CityPoor background ratingLow recruit training scorePoor probationary evaluation
Z. Officers most likely to incur charges of harassment (falsearrest, protested summons, illegal search, illegal detention,etc.) had the following characteristics:
Whites BlacksNo history of prior arrest No history of prior arrestHistory of civil count Employment disciplinary
appearances recordMilitary disciplinary record
3. Officers most likely to incur civilian complaints had thefollowing characteristics:
Whites BlacksYoung at appointment Low I. 0,Non-college graduate Many appearances in civilMilitary disciplinary record court
Military disciplinary record
4. Offic..-rs most likely to advance through the civil servicepromotional system to the ranks of sergeant, lieutenant, andcaptain had the following characteristics:
Younger at appointmentHigh I. Q.More likely to be singleCollege educatedHigher civil sorsi.-,e scoresHigher recruit training scoresMore likely to be an expert marksman
37
Table 2-1, continued
5. Officers most likely to advance to the rank of detective,which is not filled through civil service examinations,had the following characteristics:
Older at appointmentAverage I. 0.More likely to be marriedNot college educatedLower civil service scoreLower recruit training scoresLess likely to be an expert marksman
college. . . The data suggest that many men who
represent the Department's view of a desirable
candidate, especially college-educated men, will
have shorter tenure than the average officer unless
the Department consciously attempts to determine
the source of dissatisfaction among such officers
and modifies its personnel policies accordingly.23
So our debate on the value of college education is clarified
by empirical research, at least for this cohort of New York City
officers. College graduates are more likely to leave the depart-
ment, but those who remain are likely to perform better than
their colleagues. ''They advanced through preferential assign-
ments and civil service promotions, they had low i:icidence of
all types of misconduct except harassment, on which they were
average, they had low sick time, and none of them had their
38
firearms removed for cause. "24 Twenty-four percent of the
non-college graduates, but only 8 percent of the college graduates,
had received a civilian complaint. On the other hand, Cohen and
Chaiken also find that men of average intelligence and no college
education do very well in such assignments as traffic duty.25
Perhaps what police departments should be considering is
substantial restructuring of career patterns and duties and
salary structures. If college graduates, for example, were
assigned to more stimulati ng, more difficult and more highly
paid posts, their turnover might be lower and their contributions
would be used where needed most.
Well executed empirical efforts such as this have a great
deal to contribute, particularly to processing and evaluating the
facts relating to police departments' own experiences with
superior, average and inferior officers. The first problem
facing a researcher is finding meaningful performance measures,
even of the departments' own criteria. The New York City
department is large enough to make a cohort study possible.
Not many others across the country would be. Yet even Cohen
and Chaiken had to restrict themselves to measures reflecting
objective extremes of performance (termination, promotion,
discipline) because other types of evaluation (supervisor's
39
ratings, for example) tend to be extremely poor and exhibit
little real variance.
A sophisticated attempt to develop a consistent performance
evaluation method utilizing paired-comparison tests of super-
visors was made by Melany E. Baehr, John E. Furcon and
Ernest C. Froernel of the Industrial Relations Center at the
University of Chicago. 26 As predictors they used 17 different
paper-and-pencil tests to form a psychological test battery as
well as background variables. "The predictors having the
strongest correlations, consistent among subjects, were
elements of background and experience derived from a Personal
History Index, a cooperativeness variable from a test of social
insight, and temperament traits of self confidence and self-
starting (positive) and demonstrativeness (negative). "27
The tests show that, apart from an average level of
intelligence, the most important personal attributes
of successful patrolmen are all related to stability- -
stability stemming from personal self-confidence and
the control of emotional impulses, stability in the
maintenance of cooperative rather than hostile or
competitive attitudes, and stability deriving from a
resistance to stress and a realistic rather than a
40
subjective and feeling-oriented approach to life.
Underlying this stability was the presence of early
assumption of family responsibilities and involvement
in family activities in the background of better patrol-
men. All of these attributes are measured by the
tests, and patrolmen who scored high on these
attributes were, in general, those who were
independently given high ratings for performance by
their supervisors. 28
Baehr, Furcon and Froemel demonstrated that there was
an "acceptably" high and statistically significant relationship
between the psychological test battery scores and independent
measures of performance. They also demonstrated that the
relationship improved substantially when white and black
patrolmen were studied separately. On the basis of these
results, separate validations for different racial groups are
recommended as a routine procedure in the selection of patrol-
1129men.
Further discussion of the general empirical literature
here would be needlessly laborious. The best attempts have
been telescoped and indicate the inherent pos.,ibilities of the
approach. 30 One must remember, in evaluating all empirical
41
attempts, however promising, that their substantive value is
critically dependent upon their chosen measure of performance.
Internal departmental measures such as those cited here are
extremely difficult to obtain and to work with. Even more
difficult, yet perhaps more significant, are external measures
of the community's evaluation of its policemen. Since community
standards are virtually impossible to articulate even qualitatively,
pessimism seems warranted about the possibility of developing
performance measures and empirical techniques which reflect
an external evaluation of the police by the community.
Secondly, these studies fail to deal with the argument that
it is the environment encountered by the recruit in the field,
his socialization by his peers, the implicit internal system of
rewards and punishments he learns, that determines his perfor-
mance, and not his personal characteristics. To study police
success, one must study the institution and not the individuals.
If this is true, then our observed relationships may be statistical
artifacts which sidestep the primary influences.
Thirdly, in a country as heterogeneous as ours, can
results from one jurisdiction be applied usefully in other
jurisdictions or does each geographical entity have to invest in
its own research`' In spite of considerable political and technical
42
difficulties, an ambitious, careful, uniform study involving a
sampling of cities should be launched. Preliminary indications
from the existing literature, however, are pessimistic on
transferability. Moreover, perhaps the job of policeman itself,
with all its inherent contradictory elements, should be broken
down functionally (i. e. , traffic, juvenile, detective, high crime
area, low crime area, etc.) to yield meaningful results.
EVIDENCE FOR OAKLAND
If transferring results from one jurisdiction to another is
hazardous, what do we know about identifying the successful
policeman in Oakland, California, the city of our primary
concern') In fact, Oakland is a well-studied city. Ruth Levy,
in early work in the field, studied 4,000 men from 1952 to 1962
from 14 California departments, including Oakland, to try to
determine empirically the difference between men who stayed
on their jobs, men who terminated as failures (fired, resignation
requested, or ineligible for re-hire), and non-failure termina-
tions.tions. She found that non-failure terminations were the best
educated, had lived in their city for the shortest length of time,
and had the highest total number of residences. The men still
in service had lived in their city the longest periods and had the
lowest number of total residences.
43
In a later study of the same 14 departments, she dealt
with 1, 765 officers hired from November 1, 1968, through
October 31, 1969, and had access to all pre-employment informa-
tion available to the departments about these men at the time they
were hired. 32 She formed discriminant equations using 6?.
variables and 1, 599 cases. The following table shows those pre-
employment factors which occurred most frequently in the dis-
criminant equations:33
Table 2-2. Levi Profiles
CurrentMany years in Calif.Less educationHas been in NavyNot a sworn police
officerNot divorced
FailureFew years in Calif.Many jobsTatooedSworn police officerDischarged from job
Non-failure
Few years in Calif.More educationNavyOlde r
She then added to the empirically derived variables a set of
"logically derived variables" stemming from her own inter-
pretation of psychological theory. 34 Combining the logically
derived variables and the empirically derived variables in a
second discriminant equation, she reached the following results
for the logically derived variables:35
44
Table 2-3. Levy Profiles
CurrentBorn in cityNo negative
recommendationsNo bad applicationNo broken homeNo low military rankNot 21 or 22No "Service to
Mankind"Not failed other p. d.
examsFew residences
Failure
Youngest siblingNegative recommend-
ationsBad applicationBroken homeLow military rank21 or 22 years oldHad GED"Service to Mankind"
Non-failure
Oldest siblingNo negative
recommendationsJuvenile penalsNo adult penalsJr. after nameMany residencesFather is sworn
peace officerFailed other p. d. examsParent deadDishonorsHas child at young age
Levy's work has been challenged by James M. Newman,
Oakland Personnel Director, and William E. Rogin, who argue
that the predictive ability of Levy's study has not been validated
and demonstrate that for Oakland alone the predictions are no
improvement over chance. 36 They also question her classifica-
tion procedures and her emphasis on avoiding failures without
considering the cost of turning away good candidates in the
attempt. I was more disturbed by Levy's failure to fully divulge
her models, analytical techniques, and the statistical significance
of her results. I also found her "logically derived variables"
intuitively unappealing and the suggestion that such factors as
sibling position or tatoos be used for selection as bizarre and
45
frightening. Nevertheless, the effort is indicative of research
possibilities and pitfalls, and relevant to Oakland.
More compelling, from our point of view, and of utmost
importance in interpreting our own work, is the information on
tenure processed by the Oakland Civil Service staff itself. The
facts in Oakland reaffirm the disturbing evidence that the most
highly sought individuals, college graduates, are also the most
likely to resign. The candidates attracted from out-of-state,
through national recruitment efforts on the assumptions of a
national professional labor market, are also most likely to
resign.
The distribution of the 588 police officers appointed to the
Oakland Police Department in the 9 years between August 1,
1962, and August 1, 1971, by educational level, follows:37
Table 2-4. Distribution of Oakland police officers byeducational level, August, 1962, to August, 1971
Number Percent
High school graduates 218 37, 07Less than 2 years college 176 2g. 932 or more years college 98 16.67College graduates 96 16. 33
588 100. 00
46
Their distribution by place of residence is:
Table 2-5. Distribution of Oakland police officers by residence,August, 1962, to August, 1971
Number Percent
Oakland area 421 71.60Other California 57 9.70Out-of-state 110 18.70
588 100. 00
The turnover data indicate that for all appointees:
Table 2-6. Turnover of Oakland police officers, August, 1962,to August, 1971
.a.v...,...=1111,Number Percent
Remain on force 419 71.25Resigned 144 24.48Fired 25 9.25
588 99.98
Breaking down the turnover data by place of residence reveals:
Table ;?,-7. Turnover of Oakland police officers by residence,August, 1962, to August, 1971
Oakland Other Calif. Out-of-stateNo. No . % No.
Remain 318 75.5 44 77. 2 57 51.8Resigned 85 20.2 11 19.3 48 43.6Fired 18 4.2 2 3.5 5 4.0
421 99.9 57 100.0 110 99.4
47
Breaking down the turnover data by educational level reveals:
Table 2-8. Turnover of Oakland police officers by educationallevel, August, 1962, to August, 1971
H. S. <2 Years +2 Years CollegeGrad College College Grad
No. % No. No. % No. %
Remain 169 77.52 121 68.75 78 79. 59 51 53. 12Resigned 36 16.51 45 25.56 19 19.38 44 45.83Fired _i3 5.96 10 5.68 1 1.02 1 1.04
218 99.99 176 99.99 98 99.99 96 99.99
Oakland actually traced the positions taken by the 50 college
graduates who left the department between 1956 and 1971. Sixty-
eight percent of them accepted their next employment within the
field of criminal justice. This finding tempers the serious impli-
cations of the high resignation rate and indicates that, in fact,
Oakland was providing specific training for higher law enforce-
ment agencies in many cases. Whether this is a rational policy
for Oakland to follow or not depends on the trade-off between the
deriv-ttive higher personnel costs and the benefits received from
high qui,lity service, even for a few years. There may also be
a public relations or prestige gain from providing men to other
agencies. 38
48
Table 2-9. Positions taken by college graduates who leftOakland police department, 1956 to 1971
CumulativeNext Position Taken Number Percent Percent
Federal law enforcementagencies 1 24 24
State law enforcementagencies 11 22 46
Other local law enforcementagencies 11 22 68
Indirectly related to lawenforcement (law or graduateschool, military officers) 8 16 84
Other occupations 7 14 98
Fired 1 2 10050
No independent attempt will be made in this study to formu-
late more sophisticated measures of successful performance.
Such an effort would involve the analytical skills of political
science and psychology and would take us well beyond the
economic focus of this study. Moreover, we have benefited
extensively from the considerable interdisciplinary research
activity now taking place in Oakland and have been able to absorb
and include its results where relevant. 39
Among the most into
the most impressionistic,
resting and original efforts, although
is the sensitive anri provocative work
of political scientist William K. Muir. 40 He views policemen
49
as political beings, designs a system of behavioral classification,
ranks a small sample of men through in-depth interviews, and
analyzes the results. The policeman is a political being because,
in the critical part of his job, he must influence people to do what
they do not want to do, and his influence is through his a t:hority,
his threat to harm. A good policeman "must have the moral
skills to reconcile these sometimes necessary political practices
with his previously felt moral obligations to be reasonable, kind,
empathetic, and creative;" .he "has to have the moral skills to
be mean opportunistically without becoming mean compulsively. "41
"At the same time he must have the philosopher's itch--a correct2..4sense of place, time and purpose."
A "professional" policeman in Muir's scheme would have
both moral skills and a philosophical itch. This is the ideal
policeman. The "mellow" would fall short because he is a
philosopher but not a moralist, not dominant enough; the "legal-
ist" would fall short because he is a moralist but not a philoso-
pher; the ''reluctant" is neither; and the "impressionable" has
the potential to be a "professional" but is still to young and in-
experienced. Muir then uses his rankings of policemen in his
theoretical schema to assess the validity of the department's
existing selection tests and possible alternative selection de-
50
vices and to inquire into the effect of police work on personal
characteristics. Of particular interest to us, he finds that 100
percent of the "professionals" and 100 percent of the "impress-
ionables" have completed at least one year of college, as opposed
to 25 percent of the "mellows," 43 percent of the "legalists," and
0 percent of the "reluctants. "43 Again an indication that some
college may make a difference in performance while a man re-
mains on a force.
We have finished, for our purposes, our brief survey of
the policeman's job, the controversies surrounding it, and the
efforts to determine the distinguishing characteristics of
successes and failures Hopefully. one has emerged
with some understanding of the critical importance of
selection criteria, the ambiguity surrounding them, and the
difficulty of empirically validating and interpreting them. We
will turn now to the economic analysis of police selection, first
to theoretical models, then to empirical results.
51
CHAPTER 2
FOOTNOTES
1. National Commission on Law Observance and Enforcement,"Report on the Police," (Washington, D, C. : U. S. Govern-ment Printing Office, 1931), p. 19.
2. President's Commission on Law Enforcement and Adminis-tration of Justice, The Challenge of Crime in a Free Society,(Washington, D. C. : U. S. Government Printing Office,February, 1967).
3. See J. Edgar Hoover, "Should You Go Into Law Enforcement? "(New York: New York Life Insurance, 1961), p. 7. An ident-ical description is given in the Encyclopedia of Careers andVocational Guidance, Revised Edition, Volume II, (Chicago:J. G. Ferguson).
4. Jerome Skolnick argues convincingly in Justice WithoutTrial (New York: Wiley, 1967) that the coupling of dangerand authority must combine to frustrate the rule of law,that they should never be permitted to coexist. Dangertypically yields self-defensive conduct, conduct that muststrain to be impulsive because danger arouses fear andanxiety so easily. Authority then becomes a resource toreduce perceived threats rather than a series of reflectivejudgments arrived at calmly. As a result, proceduralrequirements become secondary in the face of circumstancesseen as threatening.
5. For example, a 1966 study of the Syracuse, New York,Police Department found the following distribution ofactivities, as cited in James 0. Wilson, Varieties of PoliceBehavior (Cambridge: Harvard University Press, 1968), p.
18:
Citizen Complaints: Syracuse Police Department
Ty 2e of Complaint Percent of Calls
Information gathering (book and check, geta report, inspect, question) 22.1
Service (accidents, illnesses, ambulance,animals, assist person, drunk person,escort vehicle, fire, power line, treedown, lost or found person or property,property damage)
Ofder maintenance (gang disturbance,family trouble, assault, fight,investigation, neighbor trouble)
Law enforcement (burglary in process,check a car, open door or window,prowler, make an arrest)
37. 5
30. 1
10. 3
52
See also Melaney E. Baehr, et al. , Psychological Assess-ment of Patrolman Qualifications in Relation to Field Per-formance (Washington, D. C. : U. S. Government PrintingOffice, November, 1968), pp. 9-11.
6. See Howard Vollmer and Donald L. Mills, Professionaliza-tion (Englewood Cliffs: Prentice 'Hall, 1966); Ernest Green-wood, "Attributes of a Profession," in Sigmund Nosow andWilliam H. Form, Man, Work and Society (New York: BasicBooks, 1962); Harold L. Wilensky, "Varieties of WorkExperience, " in Henry Borow, Man in a World at Work(Boston: Houghton Mifflin, 1964), pp. 125-154: and JosephW. Garbarino, "Professional Negotiations in Education, "Industrial Relations, Vol. 7, No. 2, February, 1968, pp.93-101.
7. See Jerome Skolnick, Justice Without Trial, pp. 235-239;Emile Durkheim, Professional Ethics and Civic Morals(Glencoe: The Free Press, 1958), p. 29; Max Rheinstein,Max Weber on Law in Economy and Societt(Cambridge:Harvard University Press, 1954), p. 350.
8. This list draws heavily on Arthur Niederhoffer, Behind theShield (Garden City: Doubleday, 1967), pp. 18-20; Presi-dent's Commission on Law Enforcement and Administrationof Justice, Task Force Report: The Police, and Challengeof Crime in a Free Societe, passim; and State of California,Commission on Peace Officer Standards and Training,California Law Enforcement Recruitment Program, pp. 1-9.California police departments, particularly the OaklandPolice Department, are often proudly cited as in the fore-front of professionalization.
53
9. Arthur Niederhoffer, op.. cit., p. 131.
10. James O. Wilson, 22. cit., pp. 29-30. See also HerveyA. Juris and Peter Feuille, Police Unionism (Lexington,Mass. : D. C. Heath, 1973), pp. 103-119 for sensitivetreatment of this issue.
11. President's Commission on Law Enforcement and Admin-istration of Justice, Task Force Report: The Police,pp. 167-175.
12. New York Times, Wednesday, August 15, 1973.
13. New York Times, Sunday, December 3, 1973.
14.. President's Commission on Law Enforcement and Admin-istration of Justice, Task Force Report: The Police,p. 168.
15. Oakland Civil Service Office.
16. See President's Commission on Law Enforcement andAdministration of Justice, Task Force Report: T_ he Police,pp. 167-175 for statistics confirming the general accuracyof the examples cited here.
17. See, for example, Gene E. Carte, "Police Representationand the Dilemma of Recruitment," Issues in Criminology,Vol. 6, No. 1, Winter, 1971, pp. 85-95.
18. Bernard Cohen and Jan M. Chaiken, Police BackgroundCharacteristics and Performance (New York: New YorkCity Rand Institute, August, 1972, p. ix.
19. Ibid. , pp. xi-xviii.
20. Ibid. , p. xviii. The statistical technique used here wasstepwise multiple linear regression, in which the mostpowerful background factor enters the regression equationfirst and explains as much of the variance of the performancemeasure as it can. This is followed by the second strongestindependent factor, and so forth. Whenever a new factor isintroduced, each of the preceding factors is held constant
54
to avoid duplication. The computer program utilized forthis analysis (Statistical Package for the Social Sciences)permits identification, by F test, of the statisticalsignificance of the reduction in variance produced by eachindependent variable. Only those predictor variables whosecontribution weresignificantly different from zero at the. 05level were retained. Thus, the regression equation for theperformance variable Y takes the form:
Y = a + b1X1 + b
2X
2+ ....... bkXk
where k is the number of predictor variables included, andthe particular variables included varies with the performancemeasure. See ibid. , p. 121.
21. Ibid. , pp. xviii-xix.
. 22. Ibid. , pp. xix-xxi.
a3. Ibid. , p. x.
24. Ibid. , p.
25. Ibid. , p. 157.
26. Melaney E. Baehr, Johv E. Furcon, Ernest C. Froemel,Psychological Assessment of Patrolman Qualifications inRelation to Field Performance (Washington, D. C. : U. S.Government Printing Office, November 5, 1968).
27. Bernard Cohen and Jan M. Chaiken, a. cit., p. 10.
28. National Institute of Law Enforcement and Criminal Justice,Law Enforcement Assistance Administration, U. S. Depart-ment of Justice, "Predicting Patrolman Performance,"(a summary blurb of the Baehr et al. study).
29. Baehr et al. , off. cit., p. 183..
30. The reader interested in pursuing the literature further isreferred to the following additional studies: Nick J.Colarelli and Saul M. Siegel, "A Method of Police PersonnelSelection, " The Journal of Criminal Law, Criminology, andPolice Science, Vol. 55, 1964, pp. 287-289; Robert J.
55
McDevit, "Situational Tests in Metropolitan PoliceSelection," The Journal of Criminal Law, Criminolou,and Police Science, Vol. 57, 1966, pp. 99-106: GilmoreSpencer and Robert Nichols, "A Study of Chicago PoliceRecruits: Validation of Selection Procedures, 'I The PoliceChief, Vol. 38, No. 6, June, 1971, pp. 50-55: Jewel E.Mullineaux, "An Evaluation of Predictors Used to SelectPatrolmen," Public Personnel Review, Vol. 16, pp. 84-86;Robert B. Mills, Robert J. McDevitt, and Sandra Tonkin,"Situational Tests in Metropolitan Police Recruit Selection, "The Journal of Criminal Law, Criminology, and PoliceScience, Vol. 57, 1966, pp. 99-104; Deborah Ann Tent andTerry Eisenberg, "The Selection and Promotion of PoliceOfficers: A Selected Review of Recent Literature," PoliceChief, Vol. 39, No. 2, February.1972, pp. 20-29: J. D.Matarazzo, "Characteristics of successful Policeman andFireman Applicants," Journal of Applied Psychology, Vol.48, 1964, pp. 123-133; J. T. Flynn and M. Peterson, "TheUse of Regression Analysis in Police Patrolman Selection, "Journal of Criminal Law, Criminology, and Police Science,Vol. 63, No. 4, 1972, pp. 564-569; John A. McAllister,"A Study of Prediction of Measurement of Police Perfor-mance," Police, Vol. 14, 1970, pp. 58-64; Robert W. Balch,"The Police Personality: Fact or Fiction," Journal ofCriminal Law, Criminolo , and Police Science, 1972, p.
106; Irving B. Geller, "Higher Education and Policemen,"Journal of Criminal Law, Criminology) and Police Science,1972, pp. 396-401; and Robert Hogan, "A Study of PoliceEffectiveness," Experimental Publication System, AmericanPsychological Association, Washington, D.. C., June 1970.
31. Ruth J. Levy, "Predicting Police Failures," Journal ofCriminal T aw, Criminology, and Police Science, 1967,Vol. 58, No. 2.
32. Ruth J. Levy, "Investigation of a Method for the Identifica-tion of the High-Risk Police Applicant," Can High RiskPolice Applicants , Institute for Local SelfGovernment, Berkeley, California, July 1971.
33. Ibid. , p. 35. A "current" here is defined as a man with atleast seven years of service.
56
34. For example, a first-born child, as compared with later-borns, tends to identify most particularly with authority,discipline, and parental prohibitions and moral values;exercises dominance and uses more physical power; has agreater tendency to conform when under group pressure andstress; and, therefore, other things equal, would be lesslikely to fail than a later-born child.
35. Ibid. , p. 35.
36. James M. Newman, William E. Rogin, William K. Hunter,Shanti Vora, "Investigation of a Method for Identificationof the High-Risk Police Applicant: An Analysis and Inter-pretation Based on Available Data," Can High Risk PoliceApplicants Be Identified? , Institute for Local Self Govern-ment, Berkeley, California, July, 1971.
37. Data supplied by Lamar M. Vaughn of the Oakland CivilService Department.
38. Lamar M. Vaughn and Allen Tenenbaum, "Oakland's Re-cruitment and Selection Methods: A Historical Overview, "James M. Newman, Assessment of Police Recruitment,Selection and Training: An Interdisciplinary Approach,Final Report: Recruitment and Selection, Berkeley,California, Education/ Re search, Inc. , February,1972.
39. See also: Robert Hogan, "Successful Oakland Police: APersonological Profile;" Christine M. Miller, "A Studyof the MMPI as a Selection Device for the Oakland PoliceDepartment:" Eric P. Sanders, "Oakland Police Sample:A Descriptive Psychographic Study:" and William K. Hunterand Thomas J. Culligan, "An Analysis of Factors to PredictWork Performance of Oakland Police Officers, " all in JamesM. Newman, Assessment of Police Recruitment, Selectionand Trainin : An Interdisci linary A..roach, Final Re sort:Recruitment and Selection, 2.2. cit.
40. William K. Muir, "Personal Characteristics that Contributeto Good Police Work, " in James M. Newman, Assessmentof Police Recruitment, Selection and Trainin : An Inter-disciplinary Approach, Final Report: Recruitment andSelection, 22. cit.
57
41. Ibid. , p. vii- 7.
42. Ibid. , p. vii-7. To explain Muir's context more fully, hesees that the policeman can give little positive inducementto the citizen to compensate him for his pains in desistingin an antisocial act, giving humiliating or inculpatoryinformation, going to jail. The policeman recurrentlyexercises the power to harm to extort cooperation. In this,he is a man of politics.
In politics, where threat rather than exchange is themode of behavior, the common sense of the everyday,nonpolitical world does not always obtain: paradoxesoccur, causing conventionally conceived "good"actions to serve counterproductive ends and conven-tionally conceived "bad" means to serve sociallyessential and desirable ends. A politician'sirrationality, nastiness, detachment and disregard(or reputation for these qualities) may be the ex-clusive means to dieter the irrationality, nastiness,eetachment and disregard of others under somecircumstairmes. Likewise for a good policeman; anda good policeman must have the moral skills toreconcile these sometimes necessary politicalpractices with his previously felt moral obligationsto be reasonable, kind, rmpathetic, and creative.He frequently encounters conflicting moral imperatives,and he must reconcile them while maintaining a complexsense of "right" and "wrong" without a loss of selfesteem or belief in civility. A good policeman has tohave the moral skills to be .mean opportunisticallywithout becoming mean compulsively.
At the same time he must have the philosopher's itch-,a "correct" sense of place, time and purpose. He mustdevelop an accurate perception of the motives of man-kind, an awareness of the trends of past and future,and a standard of success in keeping within both thelimits of his individual capacities and his impulse forpublic service. This social, historical, and ethicalperspective compensated for the unique distortions ofinformation which befall a man in power -- distortionsresulting from a steady diet of life's pathologies, fromthe organizational necessities of maintaining intensely
58
high morals (to overcome fear, hardship, and frustra-tion), and from the authority of office (which scaresthe timid citizen and attracts the obsequious one).
43. Ibid. , pp. vii-81.
59
CHAPTER 3
THE THEORETICAL FRAMEWORK
How well is the labor market performing in equilibrating
the demand for policemen by police departments with the supply
offered by individuals? How can the problem of persistent
"shortages" of "qullified" applicants be explained? Are wages too
low, are qualifications too high, or is the problem even more
complicated? What role is played by the institutional adjustment
mechanisms in the labor market, the recruitment, selection and
training processes? Is the market distorted by non-economic
factors" Is the market characterized by static monopsony, by
dynamic monopsony, and by problems of imperfect and uncertain
information? This chapter reviews and assesses the contribu-
tions of economic reasoning to understanding the problems posed
by the labor market for policemen.
THE SUPPLY OF POTENTIAL POLICEMEN
Calculation of Present Value of Expected Lifetime Earnin s
An individual considering becoming a policeman may
compare the present value of his expected lifetime earnings as
a policeman with the present value of his expected lifetime
earnings in his next best alternative, assuming the psychic
6o
components of the jobs are equal. If he becomes a policeman,
he may expect to receive a stream of earnings, Wri , WP2...
WPn over periods 1, 2, n, respectively. If he does not
become a policeman, his best alternative is a stream of
earnings WAI, WAz... WAR. The returns to becoming a police-
man would be the present value of the sum over periods 1 to n
of police earnings; the returns to not becoming a policeman
would be the present value of the sum of alternative earnings.
In deciding whether or not to become a policeman, the individual
would compare the expected present value of police work
1) VP = .2; -1--j-WP.j
1 + r)
with the expected present value of his next best alternative
nWA'L) VA = (1 + r)J
jr-1
If VP - VA, he would choose VP; if VP VA, he would choose
VA; if VP = VA, he would be indifferent between them.1
Since we are concerned here with expected lifetime
earnings, a term may be added to VP including the expected
earnings in future jobs should he quit or retire from the force.
His future job options should be improved because of his police
experience. He would also be concerned with evaluating whether
61
his expected lifetime as a policeman would be shorter than in
alternative occupations.
Age-Earnings Profile
Occupations may differ not only in current earnings, but
also in the temporal distribution of the expected earnings or the
age-earnings profile. For example, one occupation may have a
long training period with low or even negative earnings (direct
and opport,nity costs of schooling) followed by increasing
earning:, which peak at age 45 to 55 and then decline dramatically.
Another occupation may have no training period and earnings
which remain level throughout the working life. Present earnings
are worth more than future earnings; the interest rate is positive.
Therefore, to determine the relative pecuniary attractiveness of
the two occupations a calculation must be made which discounts
for futurity. It is the present value of the net returns for each
occupation that must be compared. In making this calculation
one would also include the relative physical exertion and other
factors which might affect the expected length of the working life,
the seasonality or cyclical sensitivity of the job, the length and
direct cost of the training period, the direct occupational
expenses required during the working life, the probability of
unemployment, and the effect of the tax structure. Different
62
individuals, given perfect information and identical tastes,
will make the same evaluation of these factors, People prefer
higher net returns to lower, other things equal. A relative wage
rate which would make the capital values in two occupations equal
can theoretically be determined.
The police job appears to have a flat age-earnings profile,
similar to that of blue-collar workers, but very different from
that of professionals. Training is done on-the-job. An Oakland
Police Department recruit in 1970 earned $926 per month during
his first year, while being trained, $960 per month during his
second year, $994 his third year, and $1, 025 his fourth year.
After his fourth year, however, he would receive only general
civil service pay increases until he retired, unless he were
promoted to sergeant. Promotional opportunities are limited to
those who can pass civil service examinations. Seventy-five
percent of the sworn personnel in Oakland are patrolmen, 15
percent are sergeants, and 10 percent are in the rank,., of
inspector, lieutenant, captain or deputy chief. Discharge from
the force is for cause only, so the policeman is protected from
swings of cyclical unemployment. The probability that he would
cont inue to earn his fourth year salary plus regular civil
service increments until his retirement (half-pay after 25 years
6.3
of service or 20 years at age 55) is high. Also included in the
actuarial comparisonwould be the value of the fringe benefits
in the police job and the alternative, the possibility of holding a
second job concurrently, and the probability of death or injury
on the job.
Shape of Supply Curve
If there were perfect competition, perfect and costless
mobility and information, and identical tastes and abilities
among individuals, everyone would evaluate identically the
relative merits of different occupations based on an actuarial
assessment, and there would be some relative wage at which
everyone would be indifferent between becoming a policeman or
doing anything else. The supply curve to an occupation would
be horizontal. It is the existence of differences among
individuals in tastes, abilities and information which leads to
differences in their evaluations of the relative wage rates
which would make two occupations equally attractive. The
supply curve to a particular occupation, therefore, will be
positively sloped.
Uncertainty
We turn now to factors which will be evaluated differently
by different people, thereby imparting a slope to the supply
64
curve. The individual decision maker will want to know not
only the present value of expected net returns, but also the
probability distribution of present values. Two occupations
may be equally attractive actuarially, but one may offer a
particular return with reasonable certainty and little probability
of variation (i. e, , the police job cited above) and the other may
offer a small chance of a very large return and a large chance
of a small return (i. e. , baseball player). Individuals will
differ in their tastes for uncertainty, and, therefore, will
assess its relative merits differently.
Tastes
The most troublesome component of the present value of
expected lifetime earnings are tastes for particular jobs.
Tastes can be expected to vary greatly among people and are
difficult or impossible to enumerate or measure. When we
define the present value of expected lifetime earnings to include
only actuarial earnings, we can also define a coefficient
measuring relative preference for police work, d. Then an
individual would choose to become a policeman if VP > (1 + d) VA,
where
VP present value of expected lifetime earnings as
a policeman
65
VA present value of expected lifetime earnings
in alternative occupation
d = coefficient measuring the taste or distaste
for police work
That is, if d is not zero, an individual would not be
indifferent between police work (P) and the alternative (A) if the
present value of expected lifetime earnings in P just equaled
those in A, but would reqiiire some compensating differential.
If, for example, he receives more psychic income in A, d is
greater than 0, and he would become a policeman only if he
receives earnings in P greater by at least d(VA) than those he
could receive in his best alternative A. In the extreme case of
a perfectly inelastic supply curve, some people would prefer
any alternative to becoming a policeman, and some people
would prefer becoming a policeman to any alternative, regard-
less of the relative pecuniary returns.
If there were no differences in tastes and abilities among
individuals, in the perfectly competitive world occupational
supply curves would be horizontal or infinitely elastic and all
wage differences would be equalizing the attractiveness of
different occupations. That is, people would be indifferent as
to which occupation they pursued. There would thus be no
66
"rents" or payments in excess of opportunity costs to any
individual. If supply curves were completely inelastic and
people chose jobs solely on nonpecuniary preferences in which
there were wide differences, relative wages would be determined
by demand, given relative supplies, and all wages would be
"rents." In the general case of a positively sloping supply
curve, returns will be equalizing only at the margin. Some
individuals will be receiving a rent in the sense that they would
be willing to work at their occupation at a lower total return.
Noncompeting groups
The exi stem e of "noncompeting groups" would also give
rise to nonequalizing wage differentials. Noncompeting groups
are formed by barriers to entry to certain occupations. These
restrictions un free choice may arise from deliberate restric-
tions on entry such as immigration laws, licensing, or trade
union power, from geographic immobilities, from differences
in ability, L from socio-economic stratification and from racial
discrimination.
The height, weight, and visual acuity standards of police
departments create a unique sort of noncompeting group peculiar
to police. William F. Danielson, Director of Personnel, City
of Berkeley, California, estimates that of 1, 031, 503 California
men between the ages of 21 and 30 in 1966, 477, 000 were 5'9"
or taller, 238, 500 were also high school graduates, and 150, 000
of these could meet other police physical, psychiatric,3
back7
ground, and mental aptitude standards. The political contro-
versies generated by these standards, and their increasingly
frequent revision, demonstrates public recognition of the
importance of these barriers to entry in the police labor market.
Nonequalizing transitional wage differentials between
occupations may also be expected from the fact that the supply
of a particular type of labor is much less elastic in the short
run than in the long run, so an change in demand would have a
much larger effect initially than ultimately. Therefore, what
is defined as an occupational wage differential depends partly
on the time period and the training for the occupation being
considered.
THE SUPPLY TO THE POLICE EMPLOYER
In a perfectly competitive world, the supply curve of a
particular kind of labor to any one employer would be infinitely
elastic or horizontal. An individual employer would be too
small to exercise independent influence on the wage, which is
set in the market at the intersection of the sum of all firms'
68
demand curves for labor of a particular kind and the sum of all
individual supply curves of labor of that kind. The employer
can hire as many workers as he wants at the market wage rate;
how many he hires is determined by his demand or value of
marginal product (VMP) curve for that labor. There should be
no persistent job vacancies, no recruiting activity by firms, no
wage discrimination, and no information costs. In a perfect
market, all potential bids and offers are known at zero cost to
every other person. Characteristics of every good need to be
known perfectly at zero cost.
However, in the real world an employer will often have
some control over his wage. He will face an upward sloping
supply curve if 1) many small firms demanding the same type
of labor are distributed over a wide geographic area; 2) workers
prefer particular jobs and places of work and/or residence;
3) information is imperfect and costs must be incurred to
inform workers of the existence of vacancies in a particular
firm. These conditions imply that to expand employment a
firm must increase wages or incur higher labor costs. The
supply curve facing him is not perfectly elastic. He has some
degree of monopsony power.
The existence of monopsony implies job vacancies at
69
equilibrium. The monopsonistic buyer will hire the number of
workers indicated by the intersection of his VMP curve with
his marginal factor cost (MFG) curve, which will be higher than
the supply curve because he will have to pay a higher wage not
only to the additional worker attracted, but also to all other
workers in his employ (assuming that he cannot discriminate
among workers). Thus, in the diagram below, he will hilt'
Q workers at wage W1, and will have vacancies at that wage
equal to Q2 -
Figure 3-1
SS
No. ofWorkers
70
Therefore, when persistent shortages appear in the real
world, monopsony is often sought as the culprit. In a competitive
framework, wages in a shortage occupation should rise relative
to wages in other occupations, thereby attracting a greater
number of workers and, ultimately, eliminating the shortage.
Indeed, David Blank and George Stigler, in their study of
the market for engineers, conclude that a "shortage exists when
the number of workers available (the supply) increases less
rapidly than the number demanded at the salaries paid in the
recent past. Then salaries will increase, and activities once
performed by engineers must now be performed by a class of
worker less well trained and less expensive. "5 In their
empirical work, they find no evidence of an increase in relative
earnings of engineers and no evidence of monopsony in the market
to distort the fu.nctioning of the wage adjustment mechanism.
They conclude, therefore, that there had been no shortage of
engine( rs.
Also, in a 1950's study of the market for engineers,
Kenneth J. Arrow and William M. Capron emphasize the role of
imperfect information and begin to formulate a simple dynamic
explanation of shortages, foreshadowing the recent more
rigorous work of Armen A. Alchian, Dale T, Mortensen,
71
Donald F. Gordon and Allan Hynes. Arrow and Capron
explain shortages--"that is, a situation in which there are
unfilled vacancies in positions where salaries are the same as
those being currently paid in others of the same type and
quality"7--by postulating a lag in the response of salaries to
demand curve shifts, a dynamic shortage. The magnitude of
the dynamic shortage depends upon the rate of increase in
demand, the reaction speed in the market, and the elasticities
of supply and demand. The lag arises from 1) the time it
takes the firm to recognize the existence of the shortage at
the current salary level; L) the time it takes to decide upon the
need for higher salaries and the number to be hired at such
salaries; and 3) the time it takes employees to realize and
react to the new alternatives available or the firm to adjust
the salaries of present employees. If there are continuing
increases in demand, then there will be a chronic shortage.
THE DEVINE MODEL
A detailed treatment of the shortage question, which
combines the roles of monopsony and imperfect information, is
developed by Eugene J Devine 8 and applied specifically to three
categories of local government employees! teachers, nurses
72
and policemen. Dev,ne argues that local governments experience
persistent shortages in these categories, though not for other
workers, because they are sufficiently large employers of
teachers, nurses and policemen to be irmonoipsonists in the labor
market, while they are competitive emptoyers for other job
categories. Moreover, a monopsonist who is unable to engage
in wage discrimination will experience job vacancies in equil-
ibrium, and will neither raise wages nor intcrease recruitment
expenditures sufficiently to eliminate them.
With imperfect. information, a monopsonist's labor supply
curve will shift downward and to the right as he increases his
recruitment expenses. Curve So in Figure 3-2, below, shows the
number of workers the firm may secure at various wage rates
if it did not engage in any recruitment activity. The Smax
curve indicates the number of persons willing to work for this
firm if they knew about its wages and employment conditions,
i. e. , if the firm were willing to incur whatever costs Were
necessary to inform these workers. OL is a "minimum avorage
outlay curve' showing the sum of the average wages and average
recruitment4costs which is the least-cost combination for
securing the number of workers indicated on the horizontal axis.
The MOL curve is the marginal outlay curve
Figure 3-2
73
No. oforkers
Q2
Devine shows that the combination of wages and recruitment
costs which is least-cost is unique for each quantity of workers.
The average wage and average' recruitment cost associated with
each number of workers is determined by an isoquant-iso-
average outlay analysis from which is derived the unique least-
cost combination of wages and recruitment costs at which each
74
quantity of workers could be obtained. The OL curve will lie
somewhere within the limits set by the So and Smax curves.
If it pays the firm to recruit in order to secure any given number
of workers such as Q1, the average total outlay must be less than
the average wage indicated by the So curve, or it would be
equally or less costly to secure 01 workers without recruiting.
On the other hand, no amount of recruitment expenditure would
enable the firm to secure 01 workers at a wage lower than that
indicated by the Smax curve, and some recruitment cost would
be necessary to get Q1 workers at that wage. The lowest average
outlay at which 01 workers could be secured is 01, as indicated
by the OL curve. Suppose this outlay consists of the sum of
some average wage Q1 -a and some average recruitment expendi-
ture 01-b (equals am). The monopsonist would be willing to hire
Qz-Qi additional workers at the equilibrium wage rate, but he
would be unwilling to pay a higher wage or to incur additional
recruitment costs to attract them. The solution presented in
Figure 3-2 is consistent, therefore, with the existence of job
vacancies.
Devine then marshalls descriptive evidence for the city
and county of Los Angeles to persuade the reader of the applica-
bility of his model to policemen, nurses and teachers. The
75
problems deriving from monopsony are intensified for police
depa:ments, for example, by the prevalence of institutional
constraints mandating parity between policemen and firemen
wages. Therefore, any increases in wages to attract additional
policemen would have to be paid not only to all policemen
already employed, but also to all firemen. The existence of
queues of qualified labor for firemen's jobs and shortages of
qualified labor for police jobs shows, moreover, that workers
are not indifferent between the two jobs at the same wages, and
adds to the difficulty of police recruiting.
David Lewin, in his complementary work on the parity
question, confirms Devine's analysis of its distorting role in
the police labor market. 9 "A policy of wage parity contributes
to these vacancy and recruitment problems by retarding the
growth of police salaries. "10 Lewin takes no position on the
monopsony explanation, but does recommend abolishing parity.
Devine's analysis should be applicable to other types of
personnel costs which may be substituted for higher wages to
increase labor supply, such as improved working conditions and
provision of on-the-job training. Yet, Devine does not deal
with the alternative implication of a monopsony analysis
developed earlier by Martin Bronfenbrenner, who points out
76
that an employer may choose not to exercise fully his monopsony
power in order to be more selective among applicants. 11 At the
monopsony wage, an employer may have to hire "nearly all
minimally qualified.workers, regardless of any special
standards of skill, experience, education, etc. which the firm
might wish to impose, and regardless also of any personnel-
office prejudices regarding age, sex, race, nationality, religion,
and so on. "12 At a higher wage, the employer will be able to
pick and choose among applicants, applying whatever hiring
standards he likes, and leaving a rejected fringe of unemployed
workers presumably composed of those least desirable to him.
Since the police shortage is a function not of insufficient
supply offered but of insufficient supply deemed e,ualified, that
is, selected, this analysis is critically important to us. Perhaps
it was ignored because it implies queues of unselected workers,
while Devine is trying to explain vacancies. Theoretically, the
two analyses can be reconciled by arguing that a wage les. than
the competitive wage but greater than the monopsony wage will
be set, therefore allowing room both for selectivity and
vacancies.
The apparent theoretical paradox of the simultaneous
existence of vacancies and queues under monopsony, combined
77
with the descriptive paradox of increasing rather than decreasing
hiring standards in the context of shortages, should serve as an
alarm that perhaps the problem is more complicated than a
simple monopsony theory allows. It is to explore one element
of this complexity tliat this thesis becomes involved in analyzing
further the selection process for policemen. Before we turn to
that, however, we would like to deal with some further theoret-
ical complexities of public employee wage determination, a
model that begins to deal theoretically with the complexities,
and, finally, some general empirical evidence.
THE PROBLEM OF PUBLIC SECTOR DECISION-MAKING
There is another problem with Devine's framework. Can
conventional economic analysis, with its profit-maximizing
basis, be applied to decision-making in the public sector,
where profit-maximization is clearly inappropriate? Public
agencies may be maximizing service subject to fixed budgetary
constraints, engaged in incremental budgetary exercises,
engaged in complex political bargaining or *.n some combination
of these. 13 How is the derived demand for labor to be defined
in a decision-making context more complex than profit-
maximization? The demand curve for labor may have to be
78
inferred indirectly through "voter expressed demands for
government services and directly through political bargaining
between governments and employee groups" rather than through
the traditional marginal revenue product curve. 14 A more
explicit consideration cf the motivation of public managers,
public workers and political processes is needed.
Devine gets around the subject by arguing that without
getting into an extended discussion of the meaning of the value
of the marginal product curve to a government agency whose
output is not priced and sold on a market and who is not profit-
maximizing, it is reasonable to assume that the VMP curve
would be negatively sloped. He postulates that the VMP curve
would show the value of the marginal produce of additional
workers to the utility-maximizing administrator of the agency.
The administrator's utility function would contain such suggested
components as the size of the agency's budget, the number of
employees supervised, his own salary, prestige, power, etc.
Nothing is said about the position of the VMP curve, or its
relation to a curve showing marginal social benefit, but it is
postulated that the curve has a negative slope because 1) the
administrator does face a budget constraint: he is competing
with other government agencies. for limited resources, and the
79
public sector is competing with the private sector; 2) he is not
immune to the consequences of his actions, and marginal social
benefits do decline as the agency expands its staff and activities;
and 3) the evidence provided by the adjustments made by local
governments to rising wages and/or vacancies is consistent
only with a VMP curve of negative slope. The evidence includes
substitution of capital for labor, reduction of the quality of the
labor force, substitution of lower-skilled employees for higher
skilled, increased recruiting activity to attract inexperienced
or cheaper labor from outside the area, on-the-job training and
improvement of working conditions.15
THE EHRENBERG ANALYSIS
A more thoughtful formulation is provided by Ronald
Ehrenberg in his superior work on the demand for state and
local government employees.16 The governmental agency is
assumed to possess a Stone-Geary type utility function which
it seeks to maximize, subject to the constraint that the total
employment budget is exhausted. That is, it seeks to maximize
nU(M) = E a
klog (f--r:)kMk ekbk)
(2.1)n
ak
> 0; ak 1; Mk-bk> 0k=1
80
subject to
where
n
E Wk Mk = Bk=1
(2. 2)
Mk is full-time equivalent state and local government
employment in the kth functional category.
ek is the factor of proportionality that converts per
capita employment into a per capita service flows
measure for the kth class.
bk is the "necessary" quantity of type k employees.
P is state population.
Wk is monthly payroll costs per employee of the kthcategory.
j is the decision-making unit.
B is the total state budget for state and local government
employees.
There are n employment categories and ak is the relative
importance of the kth type of government service. in
the decision-maker's utility function.
Maximizing (2.1) subject to (2. 2) yields employment
demand functions of the form
-W
kbka .
kW.T) =1 P
M. b.
j= 1, 2... ,n
(2, 3)
81
The notion of incremental budgeting is incorporated into
the model by assuming that the "necessary" level of employment
for a category is a function of the lagged employment level in
that category.
t-1 < <bk ak
Mk , 0 - a k - 1; k = 1, 2, , n (2. 4)
If ak = 1, employment decisions are strictly incremental; if
ak
= 0, the entire employment portfolio is a decision variable.
Intermediate values indicate that functional employment levels
can be cut to some extent.
Since tastes differ across states or within a state over
time, the relative importance of each type of government service
in the decision-makers' utility function (a ) is assumed to be a
function of a set of sociodemographic variables reflecting age
distribution (which would affect the relative importance of
educational, health and public welfare services), population
density, proportion of the population living in SMSA's, etc.
ak Vok k(X.ist)(2. 5)
The X. are those socio-demographic variables in state s inist
year t which are expected to influence the relative ma rginal
utility of the ith employment class. Then combining (2. 3)
through (2. C) and taking logs we have:
M. a.M. t -1log - = log (Voj) + L Vij log (Xist)
i=1(2. 6)
n
- log (W.) + log -Wkak M
k= 1,2, ..., n17
k=1
t -1
k
82
On the supply side, an infinitely elastic supply of labor to
each category at a predetermined wage rate has been assumed.
If the government employer does face positively sloped labor
supply curves, the estimated wage coefficients will be biased in
a positive direction. The estimates based on the previous
equations would therefore understate (in absolute terms) the
wage elasticity of demand for public employees.
Ehrenberg then attacks directly our fundamental question,
the explanation of "substantial, persistent, and wide-spread job
vacancies or shortages among certain classes of urban govern-
ment employees. "18 The model he presents is consistent with
the one we have just discussed if we assume that the recruitment
division for each specific agency is separated from the decision-.
makers who choose the authorized employment levels for all
the functional categories. The overall decision-makers were
assumed to take the quality level for each class as fixed, while
83
the recruitment division is assumed to have knowledge of the
quality distribution of applicants. 19
Ehrenberg's analysis will show that the existence of
persistent vacancy rates can be explained independently of
the existence of monopsony. Moreover, he will resolve our
apparent contradiction between vacancies and queues by showing
that vacancies are "compatible with there being a greater
number of appli-:ants than there are positions, at a wage rate
that is predetermined through either a legislative process or
collective bargaining. In particular, if applicants vary in
quality, then under certain conditions a rational government
employer will choose to employ fewer employees than his
predetermined authorized employment level. That is, the
employer will choose an equilibrium positive vacancy rate. "20
Moreover, contrary to the implications of the monopsony model,
increasing wage rates may actually increase rather than
decrease observed vacancy rates. Because this work deals so
directly with our central concerns, we will present it in
greater detail. 21
It is assumed that a government agency is allocated,
through a political process or by some other means, an
authorized employment level and a fixed wage rate per employee.
84
At this exogenously determined wage, the agency is assumed
to attract more potential applicants than it has authorized
vacant positions. The decision-makers then have to decide
whether or not to hire an applicant as he presents himself. One
way of doing this is to choose a set of minimum hiring standards
or a minimum acceptable quality level so as to maximize the
expected net gain to society. The model I s restricted to a
single period so that labor turnover and the possibility of
altering the quality distribution of applicants by either recruiting
or training expenditures can be ignored.
Applicants are assumed to be heterogeneous with qi
(where 0 qi - 1) being the expected service flow from the
.th applicant. The expected quality of each applicant is
assumed to be known once he presents himself, but the flow of
applicants to the recruiting office is a random variable with
distribution f(q). The minimum acceptable quality level is
designated as q*. All applicants exceeding this minimum hiring
standard will be employed; all others will be rejected. The
expected quality of the accepted and rejected applicants is then
given by A(q*) and .R(q*) respectively, where,
A(q*)f (q )
fqL q f (q) dq(2.14)"
fclo* o f (q) dqR (q*)
Ci) 4
(2.15)
Let W denote the specified wage rate and M denote the
authorized employment level given by the political process.
Let v denote the value per unit of the public service to society.
Let E be the total number of applicants to the agency. This
number is assumed to be greater than the number of available
positions, the entire authorized employment level (M). Given
q*, the number of applicants accepted is then
N = n(q*) E
whe re
i*
1
n(q*) = f(q)dqq(2.16)
85
An increase in the minimum acceptable quality level
increases the expected quality of both the accepted and rejected
applicants and decreases the proportion of applicants that are
accepted.
(A' , R' > 0, n' < 0)
The expected value of the net gain to society from employ-
ing this particular class of public employees when positive
vacancies exist is assumed to be given by
V = vA(q*)n(q*)E - Wn(q*)E - v1M-n(q*)E1 R(q*) (2.17)
86
The first term represents the value of the public service- -
the expected number of "quality-units" of public employees
multiplied by the value per unit. The second term represents
the direct cost--the expected wage bill. The last term represents
the opportunity cost of having vacancies--the expected loss of
public service due to the existence of vacancies. This is
assumed equal to the expected number of vacancies multiplied by
the average quality level of rejected applicants anti the value per
quality-unit of the service. It is assumed that committed funds
to pay for vacant positions sit idle. If they are used for other
purposes, the value of those uses should be added onto the
objective function.
Assuming that an interior solution exists, the first order
necessary condition to maximize (2.17) requires that the minimum
acceptable quality level (q*) be chosen to satisfy
vE [A(Oni(q9+ A' (e)n(q*)1 - WEni(e) -
v LM - n(q9Ei R' (q*) + vnt(q9ER(q*) = 0 (2. Its)
The interpretation of this condition is straightforward. An
increase in the minimum acceptable quality level has three effects.
First, it results in a smaller expected level of employment, but
increased average quality of those accepted. Secondly, by
87
decreasing the expected employment level, it decreases the
expected direct wage cost. Finally, since it causes an increase
in both the expected number of vacancies and the average quality
of rejected applicants, it increases the opportunity cost of the
potential service foregone. The agency must choose the minimum
acceptable quality level such that the net impact of these three
effects is to maximize the expected value of the net gain to
society.
Totally differentiating (2. 18), and making use of the
second-order necessary condition for (2. 18) to have a relative
maximum, we find that
a q*/ a v , a cocia M <0; a cexia w, a qqa E> 0 (2. 19)
That is, an increase in the average wage of public employees or
in the number of applicants, and a decrease in the value of the
public service or the authorized employment level, all :.mplv,
ceteris paribus, an increase in the minimum acceptable quality
level.
The minimum. acceptable quality level (q*) is not directly
observable. In practice it may consist of a vector of hiring
standards such as education levels, physical and mental charac-
teristics and health, test scores, and even residence require-
meats. Indeed, it is with defining the minimum acceptable
quality level, the vector of hiring standards, for one particular
shortage occupation, policemen, and one particular city, Oakland,
California, that the empirical focus of this thesis will be
concerned.
The ratio of actual to authorized employment (one minus the
vacancy rate) is directly observable.
(2. 20)
Differentiating (2. 20) with respect to each parameter then yields
directly
a Ev/ alAr <0; a E*/ av> 0: apq a E, aE*0 m 0 (2. 21)
Either an increase in the average wage of public employees
or a decrease in the value of a unit of the service to society will
decrease the expected ratio of actual to authorized employment,
and consequently increase the expected vacancy rate. Thus,
increasing the average wage of public employees to reduce the
number of vacancies would be a self-defeating policy. The
impact of a change in the number of applicants is in general
ambiguous, because an increase in the number of applicants
causes a reduction in the expected proportion of applicants that
are accepted. Similarly, a change in the authorized employment
89
level has an ambiguous impact. If applicants are distributed
with a uniform quality distribution, however, then an increase
in the number of applicants or a decrease in the authorized
employment level can be shown to unambiguously decrease the
vacancy rate.
If the government agency has monopsony power so that
its applicant flow will increase when there is an increase in its
average wage relative to the alternative wages which its
potential applicants perceive, the implications of the model
must be modified. The increase in the flow of applicants can
be interpreted either as an increase in the number of applicants
or as an upward shift in the quality distribution of applicants.
If the increase in the number of "higher" quality applicants is
exactly offset by a decrease in the previously lowest quality
applicants who withdraw upon observing the higher minimum
standards that have accompanied a wage increase, an upward
shift in the quality distribution alone would occur. It can be
shown that an upward shift in a uniform distribution of
applicants by quality will cause a decrease in the vacancy rate.
Indeed, here is the hope implicit in the mind of police chiefs
who advocate increasing hiring standards in the context of
shortages.
90
Thus, an increase in the average wage of public employees
will have, in this special case, one positive and two negative
impacts on the equilibrium vacancy rate. Without numerical
estimates of the magnitude of these offsetting impacts, there
should be no presumption that increasing the wages of public
employees would prove to be an effective method of reducing the
number of job vacancies in public employment.
THEORY AND REALITY
I have detailed Ehrenberg's analysis here because its
contribution to understanding our problem is outstanding. He
provides a rational theoretical underpinning, based on a
utility-maximizing decision model, for the existence of equili-
brium vacancy rates in the context of a queue of applicants and
for the possibility of raising hiring standards in the context of
shortages. He shows that increasing wage rates may not be
effective. He underscores the importance of understanding the
components of the minimum acceptable quality level, the vector
of hiring standards on which our work will attempt to cast some
light. Most importantly, his explanation is divorced from the
crutch of monopsony.
The monopsony explanation is particularly weak when
applied to policemen, whatever its validity may be for teachers
91
and nurses, both of which constitute a licensed profession.
(Shortages of teachers, however, have certainly disappeared.)
For, as we discussed in Chapter 2, the police job is amorphous,
the standards are undefined. How can an employer of ambiguous
skills and qualities be a monopsonist? A policeman becomes a
policeman not by acquiring any particular set of skills and
training (as teachers and nurses do), but by being accepted for
employment by a police departmert. Police applicants come
from a wide range of alternative jobs, spanning almost every
other occupational category. An employer who is not a large
buyer of a unique and specified type of labor, but rather seeks
certain ambiguous qualities found among people spanning the
occupational spectrum, cannot reasonably be considered a
monopsonist. Surely there are countless competitors for the
bundle of qualities, particularly good character and trainability,
that police departments seek.
Devine rationalizes this by claiming that, even for police-
men, "educational or other qualifications are such that employees
are drawn from a relatively small part of the total labor force;
none are drawn from the unskilled or semi skilled segments of
the labor force. H23 This assertion, for policemen, is just not
true,
92
Indeed, our data show just the reverse. If we classify the
applicants to the Oakland Police Department from 1965 to 1971 by
the occupational category of their most recent employment prior
to applying (usually the job held at the time of application), we
find that over two-thirds were either in the Army or in operative,
service and laborer jobs, 8 percent were craftsmen, about 17
percent were in clerical and sales jobs, and less than 9 percent
were in the professional and technical or managerial and pro-
i..rictor categories.
The tables on the following pages show the occupational
classification of applicants to the Oakland Police Department
between 1965 and 1971. Table 3-1 shows the occupational cate-
gories of the first four jobs listed on the employment history
section of the employment application. Job 1 is the job held
immediately prior to or concurrent with applying, Job 2 is the
ioh held before that, Job 3 the job held before that, etc. Many
applicants have a work history consisting of only one or two
jobs, so the number reported declines considerably as we move
towards Job 4.
Table 3-2 shows the classification of Job 1 separately for
each year to pick up any distortion of the aggregate table by
phenomena peculiar to one or two years. The unusually high
Tab
le 3
-1.
Job
App
lican
ts to
Oak
land
Pol
ice
Dep
artm
ent,
1965
-197
1, b
y oc
cupa
tiona
l cat
egor
y
Job
1Jo
b 2
Job
3Jo
b 4
Num
ber
Perc
ent N
umbe
r Pe
rcen
tN
umbe
rPe
rcen
tN
umbe
rPe
rcen
t
Prof
essi
onal
, tec
hnic
al a
ndki
ndre
d12
55.
3511
05.
5561
4.37
264
20
Man
ager
s, o
ffic
ials
,pr
opri
etor
s80
3.42
613.
0828
2.01
91.
45
Cle
rica
l and
kin
dred
260
11.1
224
212
.22
221
15.8
485
13. 7
3
Sale
s13
3.)
- 69
141
7.12
795.
6645
7.27
Cra
ftsm
en, f
orem
en a
ndki
ndre
d19
18.
1711
96.
0181
5.81
284.
52
Ope
rativ
es a
nd k
indr
ed46
820
.02
437
22.0
630
621
.94
114
18.4
2
Serv
ice
423
18.0
927
113
.68
204
14.6
280
12.9
2
Farm
labo
rers
, for
emen
and
man
ager
s4
.17
15.7
63
.22
71.
13
Lab
orer
s15
96.
8019
69.
8915
611
.18
7411
.95
Arm
y, N
avy,
Air
For
ce49
521
.17
389
19.6
425
618
.35
151
24.3
9
Tot
als
2338
1981
1395
619
Sour
ce: T
he d
ata
from
whi
ch th
is ta
ble
isco
nstr
ucte
d ca
me
from
em
ploy
men
t app
licat
ions
subm
itted
to th
e O
akla
nd C
ivil
Serv
ice
Off
ice
from
Janu
ary,
196
5, to
Jun
e, 1
971,
for
the
job
ofT
aui
Tab
le 3
-2.
Job
appl
ican
ts to
Oak
land
Pol
ice
Dep
artm
ent,
1965
-197
1, b
y oc
cupa
tiona
l cat
egor
y,fo
r Jo
b 1,
by
year
s
1965
1966
1967
1968
Num
ber
Perc
ent N
umbe
r Pe
rcen
t Num
ber
Perc
ent
Num
ber
Perc
ent
Prof
essi
onal
, tec
hnic
al a
ndki
ndre
d15
4.78
102.
37
164.
4223
4. 7
7
Man
ager
s, o
ffic
ials
,pr
opri
etor
s15
4. 7
813
3. 0
816
4. 4
218
3. 7
3
Cle
rica
l and
kin
dred
4012
. 74
5513
. 03
3710
. 22
438.
92
Sale
s19
6.05
286.
6423
6.35
214.
36
Cra
ftsm
en, f
orem
en, a
ndki
ndre
d33
10. 5
114
3. 3
234
9. 3
934
7. 0
5
Ope
rativ
e an
d ki
ndre
d70
22. 2
910
925
. 83
5615
. 47
101
20. 9
5
Serv
ice
4514
. 33
7818
. 48
8022
. 10
7014
. 52
Farm
labo
rers
, for
emen
ane
man
ager
s0
0.00
1.
241
.28
00.
0
Lab
orer
s25
7. 9
626
6. 1
627
7. 4
632
6. 6
4
Arm
y, N
avy,
Air
For
ce52
16. 5
688
20. 8
572
19. 8
914
029
05
Tot
als
314
422
362
482
Tab
le 3
-2, c
ontin
ued
1969
1970
firs
t 6 m
onth
s19
71
Num
ber
Perc
ent
Num
ber
Perc
ent
Num
ber
Perc
ent
Prof
essi
onal
, tec
hnic
al a
ndki
ndre
d10
3.41
144.
8137
21.2
6
Man
ager
s, o
ffic
ials
, pro
prie
tors
51.
71
103.
44
31.
72
Cle
rica
l and
kin
dred
3812
. 97
3110
. 65
169.
20
Sale
s15
5. 1
214
4.81
137.
47
Cra
ftsm
en, f
orem
en, a
ndki
ndre
d27
9. 2
239
13. 4
010
5.75
Ope
rativ
es a
nd k
indr
ed51
17.4
152
17. 8
729
16.6
7
Serv
ice
5619
. 11
6020
.62
3419
. 54
Farm
labo
rers
, for
emen
and
man
ager
s0
0. 0
02
.69
0e.
Oc!
Lab
orer
s20
6. 8
318
6. 1
911
6. 3
2
Arm
y, N
avy,
Air
For
ce71
24. 2
351
17. 5
321
12. 0
7
Tot
als
Z`3
329
117
4
Sour
ce: T
he d
ata
from
whi
ch th
is ta
ble.
: is
cons
truc
ted
cam
efr
om e
mpl
oym
ent a
pplic
atio
nssu
bmitt
edto
the
Oak
land
Civ
il Se
rvic
e O
ffic
e fr
omJa
nuar
y, 1
965,
to J
une,
197
1, f
or th
e jo
b of
,c)
polic
eman
.t.1
1
96
number of apphcants trorn the Army, 29 percent, in 1968, for
example, is explained by the Department of Defense 90-day
"early-out" program fo-! men entering law enforcement,
instituted in late 1967, in response to which Oakland developed
an aggressive recruitment campaign. The other really out-
standing phenomenon of the table is the marked increase from
less than 5 percent in all Other years to more than 20 percent
in the first 6 months of 1971 of men from professional and
technical occupations. This probably reflects the increasingly
difficult labor market for professionals (for example, teachers)
in the Bay Area.
In any case, the monopsony hypothesis seems substantially
useless when applied to the occupational context of the police job.
EMPIRICAL ISSUES AND ANSWERS
There have been several recent general empirical
attempts to estimate the determinants of municipal employee
wages and the elasticity of demand. In theory, one would
expect an inelastic demand for public employees. The demand
for government services is likely to be price inelastic, without
competing close substitutes. If unionized, there would be little
pressure from nonunion sectors either in the product market or
in the labor market. There is little scope for factor substitution
97
of capital for labor. The expenditures on any particular
functional category are a small part of total municipal expendi-
tures. Moreover, in the public sector political forces will
reinforce the inelasticity implied by economic ones, providing
optimal conditions for union growth and power. 24
The empirical evidence tends to bear out demand inelasti-
city. Ehrenberg estimates the demand for state and local
government employees, utilizing a time series (1958-1969) of
cross-sections with annual individual state data as the units of
observation, and finds significantly negative wage elasticities
of demand for almost all functional categories. Therefore, the
disemployment effects of real wage increases would be quite
low. His basic reduced form equation has per capita full-time
equivalent employees of category j as his dependent variable,
and includes as independent variables 1) average monthly pay-
roll cost per man, 2) the ratio of 1) to the average monthly
earnings of manufacturing production workers, 3) per capita
federal grants, 4) per capita personal income, and 5) those
sociodemographic variables which influence the demand for
class j employees. 25
Roger W. Schmenner offers some empirical work on the
determinants of teacher, police-fire, and a collection of common
98
function employee wage rates, using data for 11 American cities
over a span of 9 years, 1962-1970. His reduced form demand
equation, with the wage rate as the dependent variable, includes
1) the opportunity wage, 2) the fraction of employees with union
membership, 3) a dummy indicating the presence of formal
collective bargaining, 4) work stoppages per area employee,
showing the area's labor force militancy, 5) real property
assessments per capita, 6) the percentage change in the property
tax rate, and 7) the ratio of city population to metropolitan area
population (indicating scope for monopsony).
He concludes that unionization has played a significant role
if raising the wages of teachers: the total effect of a 0 to 100
percent membership increase and the imposition of formal
collective bargaining procedures is between 12 and .4 percent.
Police and firemen seem to benefit from complete membership
increases by about 15 percent, but the impact of formal collective
bargaining is unclear. Common function employees do not seem
to benefit from unionization, but results here are hampered by
significant data limitations.
What comparable private sector workers receive is the
most important determinant of local public sector wages, as one
would expect. Schmenner strongly agrees with our conclusion
99
that the monopsony argument could not be applicable to policemen.
"The expected positive sign and significance of the city population/
SMSA population term resoundingly destroys any notion that the
city can exercise monopsonistic power over police-fire wages. "27
He is making the point that larger cities do not have monopsony
power, but compete with smaller cities in their areas for police
labor. He does not, however, deal with the possibility of
potential monopsony power left unexercised for selection leeway,
or the problem of the definition of the police occupation. He
hypothesizes that the finding that the dominance of the central
city seems to increase salaries may indicate differences in
"tastes" among different cities or of wage differentials to compen-
sate for big city duty. Or, we might add, of higher selection
standards. For teachers, by the way, the city seems to wield
some monoptonistic power in reducing wages.
As for the other variables, pr:vate-sector labor militance,
as reflected in the number of area work stoppages, appears
important in general. The variables treating the financial and
administrative aspects of the city are of lesser importance.
James L. Freund's regression analysis demonstrates that
increases in the average weekly earnings of municipal employees
between 1965 and 1971 were signif:cantly related to inter-city
100
differences in market factors such as unemployment rates, the
pace of increases of nongovernmental wages, and the demand for
municipal workers as measured by changes in city expenditures.
Unionism, measured by the extent of organization, incidence of
strikes, and legal status of political activity by city employees,
exerted only a weak influence .28
Orley Ashenfelter's study of the effect of unionization on
firemen's wages concludes that in 1966 unionized firemen had
wages between 6 and 16 percent above those of nonunion, and
average annual duty hours of unionized firemen were from 3 to
9 percent below that of nonunion firemen.29 Ronald Ehrenberg,
also studying firemen, finds that in cities with union contracts
in 1969, hourly wages of fire fighters average from 2 to 18
percent higher than in cities without union contracts, divided
into a decrease in annual hours of between 2 and 9 percent and
an increase in annual earnings from 0 to 9 percent. 30
The contention that the demand for police services is
inelastic appears strong. It has been supported by the
empirical evidence to date. The effect, historical and
potential, of unionization, however, remains unclear, and
promises to be more elusive to measurement. For an excellent
institution discussion, see Police Unionism by Hervey A. Juris
31and Peter Feuille.
101
THE RELEVANCE OF THE THEORY
Ehrenberg gives us a formal, rigorous decision model,
stemming from a utility-based demand function for government
services and moving to derived demand functions for each
category of labor. He then looks at the problem before a
particular functional department, say police, facing an
authorized employment level and a wage rate per employee
determined and fixed elsewhere in the political process. At the
exogenously determined wage, the decision maker attracts more
potential applicants of heterogeneous quality than he has positions
to fill. He then faces the problem of which applicants to accept.
One approach to this problem is to define a minimum
acceptable quality level, say through a set of minimum hiring
standards, and then to choose those applicants who meet the
standards and reject all others. Ehrenberg concentrates on how
to express this choice in the context of his formal, utility-based
decision model. He explores how to maximize the expected
value of the net gain to society from employing this particular
class of public employees. His model allows him to show
1) the formal conditions which must be satisfied by the minimum
acceptable quality level in order to maximize the expected value
of the net gain to society; 2) the theoretical effect on the
102
minimum acceptable quality level of a change in wages, number
of applicants, authorized employment level, or value of the
public service: and 3) the theoretical effect on the expected
vacancy rate of a change in wages, number of applicants,
authorized employment level, or value of the public service.
EhrenberCs model states the problem faced, in fact,
by real world s:ate and local government civil service agencies:
given authorized employment and wage levels, who shall be
hired to fill vacant positions? The rigor of his analysis also
highlights the complex interaction between wages, hiring
standards, authorized employment levels, and vacancy rates,
a complexity too often obscured by the simpler, classical model
which ignores the selection process and implies a more direct
relationship between wage levels and vacancies, previously
disturbed only in a monopsony framework.
Yet, Ehrenberg casts no substantive light on the minimum
acceptable quality level itself, which he characterizes as a
"vector of hiring standards such as education le%els, physical
and mental characteristics and health, test scores, and even
resident requirements. "32 In fact, he dismisses it as "not
33directly observable. However, the specific vector of hiring
standards chosen by an agency molds the economic character-
103
istics of its labor force and restricts the relevant labor supply.
It is in determining these standards that the decision-maker's
pragmatic problem begins, and it is here, and in the related
areas of recruitment and training, which Ehrenberg's single-
period framework must omit, that he has real discretion to
make choices.
To aid decision - makers and public policy formulation,
specific vectors of hiring standards must be observed and their
economic consequences must be traced and analyzed. In the
policy framework of the 1970's, with its consciousness of the
discriminatory effect of past hiring standards and the need for
affirmative action, it is as important to understand the economic
conaequences of specific sets of hiring standards as to analyze
demand elasticity, wages, and wage changes. Moreover, the
definition of the desired minimum acceptable quality level of
the labor supply may be as responsible for labor "shortage"
complaints as wages.
To observe directly a vector of hiring standards, one
must do just that, observe directly. We propose to study one
agency, the Police Department of the City of Oakland, California.
We will observe, define and analyze the vector of hiring stand-
ards used for the selection of Oakland policemen from 1965 to
104
1971, and its economic content and consequences. While the
details and results of the study are applicable only to Oakland
and to the time period defined, and while Oakland is unique in
some .respects, it should be clear that the general character-
istics of the selection process discussed is by no means unique
to Oakland, but represents a prototype of the process used in
police departments, and other civil service agencies as well,
throughout the country, well back into the 1920's at least.
Against this analysis, the challenges and controversies
of 1971 to 1974 can be better understood. The details of an
appropriate and nondiscriminatory vector of hiring standards
for police departments and other agencies are now in flux.
Some old standards, challenged in the courts under Title VII
of the Civil Rights Act of 1964 and its emerging case law, are
being judged inappropriate or discriminatory. New standards,
to replace them, are emerging slowly and amid great contro-
versy. In a few years, perhaps one will be able to compare the
economic consequences of the old and the new.
We would like to begin this task by tackling the problem
of observing vectors of hiring standards and analyzing their
economic effects. We will use the Oakland Police Department
in the 196 5 to 1971 period, and describe changes made up through
July, 1974, as we proceed.
105
CHAPTER 3
FOOTNOTES
1. The treatment in this section is based on Milton Friedman.Price Theory: A Provisional Text (Chicago: Aldine, 1962),pp. 211-225.
2. It may be difficult to distinguish differences in ability fromdifferences in tastes. For example, does the deep-seadiver reap a reward for the scarce ability of being willingto work under water in dangerous circumstances or for thenonpecuniary disadvantages of the trade.
3. William F. Danielson, Police Compensation (a reportsubmitted to the President's Commission on Law Enforce-ment avri Administration of Justice, 1967), pp. 30-32.
4. Armen A. Alchian, "Information Costs, Pricing andResource Unemployment, " in Edmund S. Phelps, Micro-economic Foundations of Emplo ment and Inflation Theory(New York: W. W. Norton, 1970), pp. 27-30.
5. David M. Blank and George Stigler, The Demand and Supplyof Scientific Personnel (New York: NBER, 1957).
6. See Armen A. Alchian, "Information Costs, Pricing, andResource Unemployment;" Dale T. Mortensen, "A Theory.of Wage and Employment Dynamics; " and Donald F. Gordonand Allan Hynes, "On the Theory of Price Dynamics," allin Edmund S. Phelps, et. al. , Microeconomic Foundationsof Employment and Inflation Theory (New York: W. W.Norton, 1970).
7. K. J. Arrow and W. M. Capron, "Dynamic Shortages andPrice Rises: The Engineer Scientist Case, " QuarterlyJournal of Economics, May 1959, p. 301.
8. See Eugene J. Devine, "Manpower Shortages in LocalGovernment Employment," Ph. D. dissertation,University of California, Los Angeles, 1969: Eugene 3.Devine, Analysis of Manpower Shortages in Local Govern-
106
rnent (New York: Praeger, 1969). See also GeorgeArchibald, "The Factor Gap and the Level of Wages, 'IEconomic Record, XXX, Novembex 1954, pp. 187-199;Richard W. Hurd, "Equilibrium Vacancies in a LaborMarket Dominated by Non-profit Firms: The 'Shortage'of Nurses, " Review of Economics and Statistics, Vol. LV,No. 2, May 1973; Stuart H. Altman, "The Structure ofNursing Education and its Impact on Supply," in HerbertE. Klarman, ed. , Empirical Studies in Health Economics(Baltimore: The Johns Hopkins Press, 1970), pp. 335-352;R. T. Deane, "Simulating an Econometric Model of theMarket for Nurses, " Ph. D. dissertation, University ofCalifornia, Los Angeles, 1971; Donald E. Yett, ''Causes andConsequences of Salary Differentials in Nursing, " Inquiry,VII, March, 1970, pp. 78-99; Donald E. Yett, "The Chronic'Shortage' of Nurses: A Public Policy Dilemma, " in HerbertE. Klarman, ed. , Empirical Studies in Health Economics(Baltimore: The John Hopkins Press, 1970), pp. 357-389.
9. David Lewin, "Wage Parity and the Supply of Police andFiremen," Industrial Relations, Vol. 12, No. 1, February,1973; David Lewin, "Wage Determination in Local Govern-ment Employment," Ph. D. dissertation, University ofCalifornia, Los Angeles, 1971.
10. David Lewin, "Wage Parity and the Supply of Police andFiremen," op: cit. , p. 82.
11. Martin Bronfenbrenner, "Potential Mono,psony in LaborMarkets," Industrial and Labor Relations Review, July,1956, pp. 577-588.
12. Ibid. , p. 581.
13. For discussions of theoretical models of public employeewage determination, see Robert J. Carlsson and James W.Robinson, "Toward a Public Employment Wage Theory,"Industrial and Labor Relations Review, Vol. 22, No. 2,January, 1969, pp. 243-248; Donald Gerwin, "CompensationDecisions in Public Organizations," Industrial Relations,Vol. 9, No. 2, February,1969, pp. 175-1841 Robert J.Carlsson and James W. Robinson, "Criticisth and Comment:Compensation Decisions in Public Organizations," Industrial
107
Relations, Vol. 9, No. 1, October, 1969, pp. 111-113: JamesA. Craft, "Toward a Public Employee Wage Theory:Comment," Industrial and Labor Relations Review, Vol.23, No. 1, October, 1969, pp. 89-95. For a politicalanalysis of the bargaining process in Oakland, see FrankJ. Thompson, "The Politics of Public Personnel Policy ina Core City," Ph. D. dissertation, University of California,Berkeley, 1973.
14. Walter Fogel and David Lewin, "Wage Determination inthe Public Sector," Industrial and Labor Relations Review,Vol. 27, No. 3, April, 1974, p. 410.
15. Eugene J. Devine, "Manpower Shortages in Local GovernmentEmployment, " 2.2; cit. , pp. 25-29.
16. Ronald G. Ehrenberg, The Demand for State and LocalGovernment Employees (Lexington, Mass. : D. C. Heath,1972). In this discussion, we will number Ehrenberg'sequations as he does for the convenience of the reader whowishes to refer to his work.
17. For a more detailed discussion, see ibid. , pp. 11-16.
18. Ibid. , p. 17.
19. Ibid. , p. 18, footnote 14.
20. Ibid. , pp. 18-19.
21. Ehrenberg himself acknowledges the debt of his work to thethree papers by Alchian, Mortensen, and Gordon and Hynesin Phelps, cited in footnote 7, and to Dale T. Mortensen,"Job Search, the Duration of Unemployment, and thePhillips Curve," American Economic Review, Vol. 60,December, 1970, pp. 847-861.
22. Ehrenberg's own equation numbers are being retained herefor the ease of the reader who refers to his work.
23. Eugene J. Devine, "Manpower Shortages in Local Govern-ment Employment, " 2.2. cit. , p. 161.
108
24. See Harry H. Wellington and Ralph K. Winter, Jr. , TheUnions and the Cities (Washington, D. C. : BrookingsInstitution, 1971), pp. 17-32.
25. Ronald G. Ehrenberg, The Demand for State and LocalGovernment Employees, op. cit. , pp. 23-138.
26. Roger W. Schmenner, "The Determination of MunicipalEmployee Wages," Review of Economics and Statistics,Vol. LV, No. 1, February, 1973, pp. 83-91.
27. Ibid. , p. 87.
28. James L. Freund, "Market and Union Influences onMunicipal Employee Wages, " Industrial and Labor RelationsReview, Vol. 27, No. 3, April, 1974, pp. 391-404.
29. Orley Ashenfelter, "The Effects of Unionization on Wagesin the Public Sector: The Case of Fi re Fighters,"Industrial and Labor Relations Review, Vol. 24, No. 2,January,1971, pp. 191-202.
30. Ronald G. Ehrenberg, "Municipal Government Structure,Unionization, and the Wages of Fire Fighters, " Industrialand Labor Relations Review, Vol. 27, No. 1, October, 1973,pp. 3 6 - 48 .
31. Hervey A. Juris and Peter Feuille, Police Unionism(Lexington, Mass.-: D.. C. Heath, 1973).
109
CHAPTER 4
THE INSTITUTIONAL 'FRAMEWORK:
SALARIES, VACANCIES, HIRING STANDARDS
THE COMPOSITION OF THE OAKLAND POLICE DEPARTMENT
The Oakland Police Department, as of October 30, 1973, has
698 sworn, uniformed members. Of these, 522, or 74.8 percent,
are partrolman; 134, or 19. 2 percent, are sergeants: 42, or 6
percent, hold ranks of inspector, lieutenant, captain, deputy chief,
or chief.
Table 4-1. Oakland Police Department members, by rank,October 30, 1973
Rank Number Percent of Department
Patrolman 522 74. 8Sergeant 134 19. 2Inspector 6 0. 9Lieutenant 22 3. 2Captain 10 1. 4Deputy Chief 3 0. 4Chief 1 0. 1
Total 698 100. 0
Source: Oakland Civil Service Office
Advancement to the positions of sergeant, lieutenant and
captain is by Civil Service examination; the inspector rank is being
110
phased o.,t by merging it with sergeant' deputy chief and chief are
Civil Service exempt positions filled by appointment. To be
eligible to take the promotional exam for sergeant, for example,
three years of patrolman experience are required. A four hour
written test of 150 multiple choice questions on supervision and
leadership, analytic ability, investigation and interrogation,
operating departments, law, and rules and regulations must be
passed with a score of at least 70 percent. An oral exam must be
passed, and points are awarded for seniority. In composing the
final eligibility list, the Written counts 60 percent, the oral 30
percent and seniority 10 percent.
Because an entire generation of officers hired right after
World War II, in 1946, 1947 and 1948 reached 25 years of service
and retired by 1973, the current Oakland force is a young one.
Less than one-third of the members have ten or more years of
service: 43. 6 percent have been hired in the last five years
(see Table 4-2).
The career mobility, or pattern of advancement of cohorts
of Oakland policemen, is shown in Table 4-3.
111
Table 4-2.. Oakland Police Department members, by number ofyears of service, October 30, 1973
Number of Yearsof Service. Number Percent of Department
25 years or more(hired 1948 or before) 48 6. 9
20 to 24 years(hired 1949 to 1953) 84 12. 0
15 to 19 years(hired 1945 to 1958) 60 8. 6
10 to 14 years(hired 1959 to 1963) 33 4. 7
5 to 9 years(hired 1964 to 1968) 169 24. 2
Less than 5 years(hired 1969 to 1973) 304 43. 6
Total 698 100. 0
Source: Oakland Civil Service Office
SALARIES
Since the 1958 passage of the Proposition "C" charter
amendment, Oakland Police and Fire Department salaries are
linked to those of manufacturing production workers in the San
Francisco-Oakland area and rise at the same rate. In 1968, the
Police Officer& Association pressured the Oakland City Council
to grant salary increases beyond those authorized by the Proposi-
tion "C" formula, arguing that "C" prescribed minimum salary
increases rather than precise amounts. Under intense pressure,
Tab
le 4
-3.
Oak
land
Pol
ice
Dep
artm
ent m
embe
rs, b
y nu
mbe
r of
yea
rs o
f se
rvic
ean
d ra
nk,
Oct
ober
30,
197
3
Num
ber
ofY
ears
of
Serv
ice
Patr
olm
anSe
rgea
ntL
ieut
enan
tO
the
rPe
rcen
tN
umbe
rof
Coh
ort
Perc
ent
Num
ber
of C
ohor
tPe
rcen
tN
umbe
rof
Coh
ort
Pe r
c -n
tN
umbe
rof
Coh
ort
25 y
ears
or
mor
e(h
ired
194
8 or
befo
re)
3164
.67
14.6
36.
37
14. 6
20 to
24
year
s(h
ired
194
9 to
195
3)53
63. 1
1619
. 07
8. 3
89.
5
15 to
19
year
s(h
ired
195
4 to
195
8)17
28.3
3355
.06
10.0
46.
6
10 to
14
year
s(h
ired
195
9 to
196
3)12
36. 4
1648
. 44
12. 1
13.
0
5 to
9 y
ears
(hir
ed 1
964
to 1
968)
111
65.7
5633
. 12
1. 2
00.
0
Les
s th
an 5
yea
rs(h
ired
196
9 to
197
3)29
898
. 06
2. 0
00.
00
0. 0
Sour
ce: O
akla
nd C
ivil
Serv
ice
Off
ice
113
five councilmen decided to show "good faith" and double the amount
Proposition "C" decreed. The courts upheld the increase.
The patrolman classification, until August, 1970, had three
salary steps: patrolman, first year; patrolman, second year;
patrolman, third year. In August, 1970, a fourth step was added:
patrolman, fourth year. The fourth step was added partly to per-
mit a decrease in the starting salary without threatening the pay
scale of incumbents and partly to widen the career ladder. Once
a patrolman reaches the fourth step, he must be promoted to
sergeant to increase his salary scale. If he is not promoted to
sergeant, his pay will increase only by yearly cost of living
increases for patrolmen, fourth year. There are no seniority or
merit increases. A patrolman on the force for 24 years will be
earning the same salary as a patrolman on the force for four years.
The monthly salary for each patrolman step, sergeant and
lieutenant from July, 1963, to July, 1974, is given in Table 4-4.
The salary differences between patrolman, first year,
patrolman, second year, patrolman, third year, patrolman,
fourth year, and sergeant are given in Table 4-5.
In other words, a patrolman's salary increases about three
and one-half percent between his first year of service and his
second year of service and between his second year and third
Tab
le 4
-4.
Mon
thly
sal
ary
of O
akla
nd P
olic
e D
epar
tmen
t mem
bers
, by
rank
, Jul
y, 1
963,
toJu
ly, 1
974
Patr
olm
anFi
rst Y
ear
Patr
olm
anSe
cond
Yea
rPa
trol
man
Thi
rd Y
ear
Patr
olm
anFo
urth
Yea
rSe
rgea
ntI
ieut
enan
t
July
, 196
3$6
2.5
$647
$667
$758
$836
July
, 196
464
666
968
978
386
4
July
, 196
566
368
670
780
388
6
July
, 196
668
671
073
183
191
7
July
, 196
771
173
675
886
195
0
July
, 196
877
380
082
493
61,
033
Aug
ust,
1968
823
852
878
992
1, 0
95
July
, 196
987
390
493
21,
053
1, 1
62
July
, 197
096
099
41,
025
1, 1
581,
344
Aug
ust,
1970
926
960
994
$1, 0
25Ju
ly, 1
971
977
1, 0
121,
048
1, 0
811,
221
1, 4
17
July
, 197
21,
058
1, 0
951,
134
1, 1
701,
322
1, 5
34
July
, 197
31,
134
1, 1
741,
215
1, 2
541,
417
1, 6
44
July
, 197
41,
201
1, 2
431,
286
1, 3
281,
527
1, 7
56
Sour
ce: O
akla
nd C
ivil
Serv
ice
Off
ice.
Tab
le 4
-5.
.D
iffe
renc
es in
sal
ary
of m
embe
rs o
f th
e O
akla
nd P
olic
e D
epar
tmen
t, by
ran
k,Ju
ly, 1
963,
to J
uly,
197
4
Patr
olm
anFi
rst Y
ear
to S
econ
d Y
ear
Patr
olm
an'
Seco
nd Y
ear
to T
hird
Yea
r
Patr
olm
anT
hird
Yea
rto
Fou
rth
Yea
r
Patr
olm
anL
ast Y
ear
to S
erge
ant
Dol
lars
Perc
ent
Dol
lars
Perc
ent
Dol
lars
Perc
ent
Dol
lars
Perc
ent
July
, 196
3$2
23.
52$2
03.
09
$91
13.6
4Ju
ly, 1
964
233.
56
203.
00
9413
. 64
July
, 196
523
3.47
213.
06
9613
.58
July
, 196
624
3. 5
021
2. 9
610
013
.68
July
, 196
725
3. 5
222
2. 9
910
313
.59
July
, 196
827
3.49
243.
00
112
13.5
9A
ugus
t, 19
6829
3. 5
226
3.05
114
12.9
8Ju
ly, 1
969
313.
55
283.
10
121
12.9
8Ju
ly, 1
970
343.
54
313.
12
133
12.9
8A
ugus
t, 1.
970
343.
6734
3. 5
4$3
13.
12
July
, 197
135
3. 5
836
3. 5
633
3. 1
314
012
. 98
July
, 197
237
3. 5
039
3. 5
637
3. 2
615
212
. 99
July
, 197
340
3. 5
341
3. 4
939
3. 2
116
313
. 00
July
, 197
442
3. 5
043
3.46
423.
27
199
13. 0
0
Sour
ce: D
ata
on w
hich
cal
cula
tions
are
bas
ed w
as s
uppl
ied
byO
akla
nd C
ivil
Serv
ice
Off
ice.
116
year, and about three and one-quarter percent between the third
year and fourth year. Then, unless he advances in rank, he
receives only cost-of-living increases, pegged to the increase
in average earnings of Bay Area manufacturing production
workers.
The annual increases in patrolman, first year, classifica-
tion salaries and the annual increases in wages of Bay Area
manufacturing production workers are shown in Table 4-6.
The relative wages of Oakland patrolman and Bay Area
production workers changed most dramatically with the double
increase won by the patrolman in 1968. The patrolman'. s starting
salary averaged slightly over 115 percent of production workers'
wages before August, 1968, and over U5 percent afterwards.
Patrolmen in their last step went from about 125 percent of
production workers' wages before August, 1968, to close to 140
percent afterwards. Sergeants' salaries were about 140 percent
of production workers' wages before August, 1968, 157 percent
afterwards. The ratio of patrolmen's salaries, first and last
year, and sergeant's salaries, to average production workers'
earnings, is presented in Table 4-7.
It should be stressed here that the figures for Oakland
police salaries and for average production worker wages are not
Tab
le 4
-6.
Ann
ual r
ate
of in
crea
se, O
akla
nd P
olic
e D
epar
tmen
t pat
rolm
an a
nd S
anFr
anci
sco-
Oak
land
man
ufac
turi
ng p
rodu
ctio
n w
orke
rs, J
uly,
196
3 to
Jul
y, 1
973
Dat
e
Oak
land
Pat
rolm
an,
Firs
t Yea
rPr
oduc
tion
Wor
kers
in M
anuf
actu
ring
San
Fran
cisc
o-O
akla
nd
Ann
ual S
alar
yPe
rcen
t Rat
eof
Inc
reas
ePe
rcen
t Rat
eA
nnua
l. W
age
of I
ncre
ase
July
, 196
2$
7, 2
72$
6, 2
03Ju
ly, 1
963
7, 5
001.
56,
457
4. 1
July
, 196
47,
752
3. 4
622
2.6
July
, 196
57,
956
2.6
6, 9
154.
4Ju
ly, 1
966
8, 2
323.
57,
253
4. 9
July
, 196
78,
532
3.6
7, 3
210.
9Ju
ly, 1
968
9, 2
768.
77,
907
8. 0
July
, 196
910
, 476
12.6
8, 4
026.
3Ju
ly, 1
970
11, 5
2010
. 08,
549
1. 7
Aug
ust,
1970
11, 1
12-3
. 58,
621
0. 8
July
, 197
111
, 724
5. 5
9, 3
338.
3Ju
ly, 1
972
12, 6
968.
310
, 131
8. 6
July
, 197
313
, 608
7. 2
10, 8
597.
2
Sour
ce: D
ata
on w
hich
cal
cula
tions
for
Oak
land
are
bas
ed w
ere
supp
lied
by O
akla
nd C
ivil
Serv
ice
Off
ice.
Dat
a on
whi
ch c
alcu
latio
ns f
or p
rodu
ctio
n w
orke
rs in
man
ufac
turi
ng,
San
Fran
cisc
o-O
akla
nd, a
re b
aser
'. w
ere
supp
lied
by S
tate
of
Cal
ifor
nia,
Dep
artm
ent
of I
ndus
tria
l Rel
atio
ns, D
ivis
ion
of L
abor
Sta
tistic
s an
d R
esea
rch,
Cal
ifor
nia
Tab
orSt
atis
tics
Bul
letin
, Jul
y, 1
962,
to J
uly,
197
0, a
nd S
tate
of
Cal
ifor
nia,
Dep
artm
ent o
fH
uman
Res
ourc
es D
evel
opm
ent,
Cal
ifor
nia
Lab
or M
arke
tB
ulle
tin, J
uly,
197
1, to
July
, 197
3.
118
Table 4-7. Ratio of C-ikland Police Department patrolman andsergeant salaries to earnings of San Francisco-Oaklandmanufacturing production workers, July, 1963, to July,1973
Patrolman, 1 Patrolman, 3 SergeantDate Average Earnings Average Earnings Average Earnings
July, 1963 1. 16 1. 24 1. 41July, 1964 1. 17 1. 25 1.42July, 1965 1. 15 1. 23 1.39July, 1966 1. 13 1. 21 1.37July, 1967 1. 16 1. 24 1.41July, 1968 1. 17 1. 25 1.42July, 1969 1. 25 1.33 1. 50
July, 1970 1.35 1.44 1.63July, 1971 1.26 1.39 1.57July, 1972 1. 26 1.39 1.57July, 1973 1. 25 1.39 1.57
Note: Ratios are calculated by dividing annual salary of OaklandPolice Department member by average annual wages ofproduction workers in manufacturing in the San Francisco-Oakland area. For example, for July, 1973, patrolman,first year, the calculation is 13,608/10, 859 equals 1. 25.
Source: Data for Oakland were supplied by the Oakland CivilService Office. Data for production workers in manufac-turing, San Francisco-Oakland, were supplied by State ofCalifornia, Department of Industrial Relations, Division ofLabor Statistics and Research, California Labor StatisticsBulletin, July, 1963, to July, 1970, and State of California,Department of Human Resources Development, CaliforniaLabor Market Bulletin, July, 1971, to July, 1973.
119
measuring the same thing. The Oakland figures are derived
simply by multiplying the monthly salary rate in effect in July
of each year by 12. They represent salary scales, not actual
earnings. For example, overtime pay earned by policemen
would not be reflected. The production worker earnings figures
are derived by multiplying average weekly earnings of San
Francisco-Oakland SMSA production workers in July by 52. They
represent actual earnings and would reflect the number of hours
worked and overtime pay as well as wage rates. July figures
were chosen in both cases because the fiscal year in Oakland
commences in July and most salary changes take effect then.
The comparability is sufficient for our rough purposes.
An even more revealing set of relative wage calculations
can be provided for the census year 1969. In Tables 3-1 and 3-2,
we presented the previous occupations of the applicants to the
Oakland Police Department between 1965 and 1971. We know,
for example, that in 1969, 3. 4 percent of the applicants came
from professional, technical and kindred jobs, 1. 7 percent from
managerial, almost 13 percent from clerical, 5.1 percent from
sales, 9. 2 percent from craftsman, 17. 4 percent from operative,
19.1 percent from service, 6. 8 percent from laborer jobs and
24. L percent from the military. The 1970 Census provides us
120
with median earnings for San Francisco-Oakland SMSA workers
in each of the civilian occupational categories.) Table 4-8
compares the annual salary of Oakland patrolman, step 1,
patrolman, step 3, and sergeant, in 1969 with the median
earnings of San Francisco-Oakland SMSA workers in other broad
occupational categories. The ratio of the police salary to median
earnings in each occupational category is presented for the total
population and for the Negro and Spanish surnamed sectors
separately.
The Oakland figures are based on the monthly salary rate
after the July, 1969, increase. Again, salary scale figures and
earnings figures are not strictly comparable. The Census earn-
ings figures not only reflect overtime, but also include earnings
from second jobs or businesses. All members of the experienced
civilian labor force in the occupation are included, whether
employed or unemployed. The police salary scale figures do not
reflect overtime or second jobs or unemployment, as the earnings
figures do. Both overtime and second job opportunities represent
important additions to primary police salaries.
Because of its interest, we report the protective service
category in more detail than the other occupational categories.
Oakland patrolmen, step 1, earn only 86 percent of firemen's
Tab
le 4
-8.
Oak
land
Pol
ice
Dep
artm
ent s
alar
ies
for
patr
olm
an, s
tep
1, p
atro
lman
, ste
p 3,
and
ser
gean
t,re
lativ
e to
med
ian
earn
ings
in o
ther
occ
upat
iona
l gro
ups,
196
9, f
or S
an-F
ranc
isco
-Oak
land
SMSA
, tot
al,
Neg
ro a
nd S
pani
sh s
urna
me
Occ
upat
iona
l Cat
egor
y
Tot
alN
egro
Span
ish
Surn
ame
Med
ian
Ear
ning
sPa
t.1Pa
t.3Sg
t.M
edia
nE
arni
ngs
Pat.1
Pat.3
Sgt.
Med
ian
Ear
ning
sPa
t.1Pa
t.3Sg
t.M
. E.
M. E
.M
. F.
M. E
.M
. F.
M. E
.M
. F.
M. F
.M
. F.
Exp
erie
nced
civ
ilian
labo
r fo
rce,
all
$ 9,
073
1. 1
5I.
23
1. 3
9$6
, 968
1. 5
01.
611.
81
$ 8,
098
1. 2
91.
38
1. 5
6
Prof
essi
onal
, tec
hnic
al11
, 784
.89
.95
1. 0
78,
126
1. 2
91.
38
1. 5
610
, 178
1. 0
31.
10
1. 2
4
Man
ager
s,ad
min
istr
ator
s12
, 992
.81
.86
.97
9, 1
501.
14
1. 2
21.
38
11, 7
48.
89.
951.
08
Sale
s9,
372
1. 1
21.
19
1.35
6, 2
811.
67
1. 7
82.
018,
796
1. 1
91.
27
1. 4
4
Cle
rica
l7,
576
1. 3
81.
481.
676,
846
1. 5
31.
631.
857,
290
1. 4
4L
53
1. 7
3
Cra
ftsm
en9,
757
1.07
1.15
1.30
7, 8
581.
33
1. 4
21.
619,
076
1. 1
51.
231.
39
Ope
rativ
es, e
xc.
tran
spor
t7,
594
1. 3
81.
47
1.66
7, 0
481.
49
1. 5
91.
797,
466
1. 4
01.
50
1.69
Tra
nspo
rt e
quip
men
top
erat
ives
8, 6
661.
21
1. 2
91.
46
7, 5
811.
38
1. 4
81.
678,
951
1. 1
71.
25
1.41
Lab
orer
s, e
xc. f
arm
6, 6
191.
58
1.69
1.91
6, 8
541.
53
1.63
1. 8
47,
161
1. 4
61.
56
1.76
Farm
ers
and
farm
man
ager
s6,
696
1. 5
61.
671.
893,
364
3. 1
13.
32
3. 7
67,
065
1.48
1. 5
81.
79
Farm
labo
rers
and
fore
men
3, 8
602,
71
2. 9
03.
27
1.63
96.
39
6. 8
27.
714,
330
2. 4
22.
58
2. 9
2Se
rvic
e, e
xc. h
ouse
hold
6, 1
831.
691.
81
2. 0
45,
656
1.85
1.98
2. 2
36,
057
1. 7
31.
85
2. 0
9
Prot
ectiv
e se
rvic
e10
, 831
.97
1.03
1. 1
76,
241
1.68
1.79
2. 0
210
, 153
1. 0
31.
10
1. 2
4
Fire
men
12, 1
7286
.92
1. 0
412
, 292
.85
.91
1.03
Guz
.rds
and
wat
chm
en5,
869
1.78
1.91
2.15
4, 7
682.
20
2. 3
52.
655,
500
1.90
2.03
2.33
Polic
emen
and
dete
ctiv
es11
, 767
.89
.95
1. 0
79,
875
1. 0
61.
13
1. 2
810
, 633
991.
05
1. 1
9
Priv
ate
hous
ehol
d2,
531
4. 1
44.
42
4. 9
91,
882
5. 5
75.
94
6.71
Tab
le 4
-8, c
ontin
ued
Not
e: T
he a
nnua
l sal
ary
for
patr
olm
an, s
tep
1 (P
at. 1
),in
Jul
y, 1
969,
was
$10
, 476
: the
ann
ual s
alar
y fo
rpa
trol
man
, ste
p 3
(Pat
. 3),
in J
uly,
196
9, w
as $
11, 1
84: t
he a
nnua
l sal
ary
for
serg
eant
in J
uly.
1969
,w
as $
12,6
36T
he -
sig
n m
eans
that
the
med
ian
earn
ings
fig
ure
was
not
pre
sent
ed b
y th
e C
ensu
s be
caus
eth
e ba
seis
sm
alle
r th
an th
e m
inim
um n
umbe
r pr
escr
ibed
for
the
sam
ple.
Sour
ce: D
ata
for
Oak
land
wer
e su
pplie
d by
the
Oak
land
Civ
il Se
rvic
e O
ffic
e.D
ata
on m
edia
n ea
rnin
gs(M
. E.)
for
the
San
Fran
cisc
o-O
akla
nd S
MSA
, by
occu
patio
n an
d ra
ce, c
ome
from
U. S
.D
epar
tmen
t of
Com
mer
ce, S
ocia
l and
Eco
nom
ic S
tatis
tics
Adm
inis
trat
ion,
Bur
eau
of th
e C
ensu
s,19
70 C
ensu
s of
Pop
ulat
ion,
Det
aile
d C
hara
cter
istic
s, C
alif
orni
a, S
ectio
n 1
(PC
(1)
-D6
Cal
if.)
, pp.
Cal
if 6
-182
3 to
6-1
828,
Tab
le 1
75, "
Occ
upat
ion
of th
e M
ale
Exp
erie
nced
Civ
ilian
Lab
orFo
rce
byE
arni
ngs
in 1
969
and
Rac
e: S
an F
ranc
isco
-Oak
land
SM
SA."
123
earnings and 89 percent of all area policemen's and detectives'
earnings; Oakland patrolmen, step 3, earn 92 percent of fire-
men's earnings and 95 percent of all area policemen's and
detective& earnings. Yet, Oakland Police Department salaries,
along with San Francisco's, are the highest in the Bay Area, as
well as in the state. We know, too, that firemen's salaries are
pegged to policemen's salaries. Median earnings will be
affected somewhat by earnings of policemen in higher ranks than
patrolman. But the apparent puzzle is best explained by the
prevalence of both overtime and second jobs or businesses
among policemen and firemen. Therefore, one would expect
overall earnings figures to exceed those based on salary scales,
as our ratios demonstrate. Firemen, because of the arrangement
of their working hours into consecutive toll day and night duty
tours and consecutive days off, have even more opportunities
for additional earnings than policemen. It is not surprising,
therefore, that the differential between Oakland police salary
scales and average earnings is even greater for firemen than
for policemen. We can infer that roughly 95 percent of the total
earnings of the average Oakland patrolman, step 3, comes from
his stated salary scale; second jobs and overtime supply the rest.
Second jobs are permitted, but must be revealed and approved by
124
the chief. It should also he stressed here that graft is not a
problem in the Oakland Police Department and illegal income can
reasonably be ignored for purposes of our work.
The only other occupational categories in Table 4-8 in
which median earnings exceed Oakland patrolman earnings are
professional and technical and managerial and administrative.
It is not surprising that such a small percent of total police
applicants was drawn from these categories when Oakland
patrolmen start at 89 percent of the median earnings of pro-
fessionals and 81 percent of the median earnings of managers in
the area, and in their third year earn 95 percent and 86 percent
respectively of professional and managerial median earnings.
Even Oakland sergeants earn only 97 percent of the median
earnings of managers, although 107 percent, now, of those of
professionals and technicians.
Similarly, it is not surprising that most of the civilian
applicants to the Oakland Police Department come from service,
operative and clerical jobs. The Oakland patrolman's starting
salary is 169 percent of the median earnings of service workers
and 138 percent of the median earnings of clerical workers and
operatives. The differentials are even greater in the laborer,
farm, and private household worker categories, but workers in
125
these categories are much less likely to meet the other require-
ments, particularly high school graduation, of the Oakland Police
Department. It should also be noted here that Oakland patrolmen
start at 178 percent of the median earnings of private guards and
watchmen.
The differential between Oakland patrolman earnings and
median earnings in other occupational categories is greater,
across all categories, for the Negro and Spanish surnamed
population than for the total population. This is because median
earnings across all occupations is lower for Negroes and Spanish
surnamed individuals. Does this imply, as was so often argued
in discussions of a volunteer army, that the Oakland Police
Department is flooded with applications from the minority
community because becoming a policeman would have relatively
more pecuniary attractiveness to a minority person than to
others? We will deal with the minority composition of the
Department, and the issues surrounding it, in detail in Chapter
6. The answer to our question is clearly no, because pecuniary
attractiveness does not tell the whole story. In fact, minority
community members clearly are more reluctant than others to
become policemen. The distaste attached to becoming a police-
man is much greater in the minority communities because of the
126
feel ing, whether ove rt or inc ipient, that minority communities
have been occupied and harrassed by alien, hostile police forces
throughot.' their history, and that joining the police force would
be joining the enemy. On the personal level, the minority
policeman may be both rejected by his community and his fellow
officers. Moreover, minorities do not see themselves on police
forces, do not feel they would be selected or accepted. To
attract more minorities, large pecuniary incentives are not
enough. The taste problem must be tackled directly, through
recruitment campaigns aimed at the minority communities,
through changes in the image of the police force, through changes
in selection techniques, and, perhaps, through more fundamental
changes in composition and definition of the police job.
We can present one more set of interesting relative income
calculations for the census year 1969. We can show the relation-
ship of Oakland police salaries to the mean income of San Fran-
cisco-Oakland area men by years of school completed, age and
race. Income figures are even less comparable to scheduled
Oakland police salaries than the earnings figures presented
previously, because income includes not only earnings, but
interest, dividends, veterans' payments, social security,
pensions, etc. , L as well. Nevertheless, the comparisons allow
Tab
le 4
-9.
Oak
land
Pol
ice
Dep
artm
ent p
atro
lman
sta
rtin
g sa
lary
rel
ativ
e to
mea
nin
com
e of
all
San
Fran
cisc
o-O
akla
nd m
ales
, 18
and
olde
r, a
nd 2
5 to
34,
in 1
969,
by
year
s of
scho
ol c
ompl
eted
and
race
All
One
to T
hree
Five
Yea
rsE
duca
tiona
lFo
ur Y
ears
Yea
rsFo
ur Y
ears
Plus
Gro
ups
Hig
h Sc
hool
Col
lege
Col
lege
Col
lege
Mea
nIn
:. om
ePa
t. 1
Mea
nIn
com
ePa
t. 1
Mea
nIn
com
ePa
t. 1
Mea
nIn
com
ePa
t. 1
Mea
nIn
com
ePa
t. 1
M. I
.M
. I.
M. I
.M
. I.
M. I
.
All
mal
es,
18 a
ndol
der
Tot
al$9
,27
41.
13
$8,
609
1.22
$8,
855
1. 1
8$1
3.23
279
$15,
568
.67
Neg
ro5,
949
1.76
6,02
61.
746,
427
1.63
8,60
61.
2211
,35
0.92
Span
ish
surn
ame
7,88
41.
33
8,02
61.
317,
939
1.32
10,
949
.96
13,
527
77
Mal
es, 2
5 to
34
Tot
al$8
,90
31.
18
$8,
736
1. 2
0$8
,76
21.
20
$10,
059
1.04
$ 9.
970
1.05
Neg
ro6,
637
1. 5
86,
719
1. 5
67,
047
1. 4
98,
576
1.22
8,20
31.
25
Span
ish
surn
ame
8,21
91.
27
8,45
71.
24
8,30
01.
26
9,20
91.
149,
747
1. 0
4
Not
e: T
he a
nnua
l sta
rtin
g sa
lary
for
Oak
land
pat
rolm
an, s
tep
1 (P
at.
1),
in J
uly,
196
9, w
as $
10, 4
76.
Sour
ce: D
ata
for
Oak
land
wer
e su
pplie
d by
the
Oak
land
Civ
ilSe
rvic
e O
ffic
e.D
ata
on m
ean
inco
me
(M. I
.) f
or th
e Sa
n Fr
anci
sco-
Oak
land
SM
SA, b
y ag
e, r
ace,
and
yea
rsof
edu
catio
n, c
ome
from
U. S
. Dep
artm
ent o
f C
omm
erce
, Soc
ial a
nd E
cono
mic
Stat
istic
s A
dmin
istr
atio
n,B
urea
u of
the
Cen
sus,
197
0 C
ensu
s of
Pop
ulat
ion,
Det
aile
dC
hara
cter
istic
s, C
alif
orni
a.Se
ctio
n 2
(PC
(1)
-D6
Cal
if. )
,pp
. Cal
if 6
-233
6 to
6-23
38, T
able
197
, "In
com
e in
196
9 of
Pers
ons
18 Y
ears
Old
and
Ove
r by
Yea
rs o
f Sc
hool
Com
plet
ed,
Age
, Rac
e an
d Se
xSa
n Fr
anci
sco-
Oak
land
SM
SA. "
Tab
le 4
-10.
Oak
land
Pol
ice
Dep
artm
ent p
atro
lman
, ste
p 3,
and
ser
gean
t,sa
lari
es r
elat
ive
to m
ean
inco
me
of a
llSa
n Fr
anci
sco-
Cak
land
mal
es, 1
8 an
d ol
der,
in 1
969,
by
year
s of
scho
ol c
ompl
eted
and
rac
e
All
EzT
uca.
tiona
lFo
ur Y
ears
Gro
ups
Hig
h Sc
hool
Mea
nPa
t.3M
ean
Pat.3
Sgt
.In
com
e M
. I. M
. I. I
ncom
e M
. I. M
. I
One
to T
hree
Yea
rsC
olle
geFo
ur Y
ears
Col
lege
Five
Yea
rsPl
usC
olle
ge
Mea
nPa
t.3 S
gt.
Mea
nPa
t.3 S
gt.
Mea
nPa
t.3 S
gt.
.In
com
e M
. 1. M
. I.
Inco
me
M. I
.M
. I.
Inco
me
M. I
.M
. I
All
mal
es,
18 a
nd o
lder
Tot
alN
egro
$9, 5,
274
949
1. 1.21 88
1. 2.36 12
$8, 6,
609
026
1.30
1. 8
61. 2.
47 10$8
, 6,85
542
71. 1.
26 741.
43
1.97
$13, 8,
232
606
.
1.85 30
.
1.95 47
$15, 11
,56
835
0.72
.98
.8
1. 1
7,88
41.
421.
608,
026
1.39
1.39
7,93
91.
411.
59
10,
949
1.02
1.15
13.
527
.83
.9
Not
e: T
he a
nnua
l sal
ary
for
Cak
land
pat
rolm
an, s
tep
3 (P
at. 3
), in
Jul
y, 1
969,
was
$11
, 184
; the
ann
ual
sala
ry f
orse
rgea
nt (
Sgt.)
in J
uly,
196
9, w
as $
12,6
36.
Sour
ce: D
ata
for
Oak
land
wer
e su
pplie
d by
the
Oak
land
Civ
il Se
rvic
e O
ffic
e.D
ata
on m
ean
inco
me
(M. I
.) f
or th
e Sa
nFr
anci
sco-
Oak
land
SM
SA, b
y ag
e, r
ace,
and
yea
rs o
f ed
ucat
ion,
com
efr
om U
. S. D
epar
tmen
t of
Com
mer
ce.
Soci
al a
nd E
cono
mic
Sta
tistic
s A
dmin
istr
atio
n,B
urea
u of
the
Cen
sus,
197
0 C
ensu
s of
Pop
ulat
ion,
Det
aile
dC
hara
cter
istic
s, C
alif
orni
a, S
ectio
n 2
(PC
(1)
-D6
Cal
if.)
, pp.
Cal
if. 6
-233
6 to
6-2
338.
Tab
le 1
97, "
Inco
me
in 1
969
of P
erso
ns 1
8 Y
ears
Old
and
Ove
rby
Yea
rs o
f Sc
hool
Com
plet
ed, A
ge, R
ace
and
Sex-
San
Fra
ncis
co-
Oak
land
SM
SA."
OD
129
us to demonstrate that Oakland police salaries are considerably
more attractive to high school graduates than to college graduates.
In fact, if we look at all San Francisco-Oakland men over the age
of 18, Oakland police starting salaries are 122 percent of the
mean income of high school graduates, 118 percent of the mean
income of men with one to three years of college, and only 79
percent of the mean income of college graduates.
Oakland Police Department applicants, however, must be
between the ages of 21 and 32. The mean income for this younger
age group is below that for all men over 18. The Census data for
18 to 24 year old men is clearly distorted by school attendance
and part-time earnings. If we take men 25 to 34 years old, we
find that the Oakland police starting salary is greater than the
mean income of high school graduates and men with one to three
years of college, and 104 percent of the mean income of college
graduates. There is clearly a substantial difference in the
pecuniary attractiveness of police work for high school and
college graduates. The data are presented in Table 4-9 and
Table 4-10.
BENEFITS
Oakland patrolmen receive benefits which add, by the
city' s estimates in May, 1971, an additional $5, 829 to an Oakland
130
Table 4-11 Total remuneration of Oakland patrolman, step 3,fiscal year July, 1970, to June, 1971
Source Dollar Amount
Salary $ 12, 300Retirement 5, 045Workmen's Compensation 386Health 180Dental 53
Uniform 165
Total $ 18, 129
Source: Oakland Civil Service Office
patrolman's third year salary. 3 His total remuneration is shown
in Table 4-11.
The retirement program has the following featur.3s:
1) One half pay after 25 years of service. One half pay
at age 55 with 20 years of service.
2) Retirement pay is 50 percent of whatever current pay
is. That is, an officer who retired in 1950 receives 50
percent of today's top pay, or $512. 50 per month.
3) Seventy-five percent pay to an officer who is retired for
service-incurred disability. Upon reaching his normal
retirement date he assumes the normal 50 percent
salary.
131
4) Seventy-tive percent pay to dependents of a retired
officer who dies as a result of service injuries. Upon
normal retirement date this amount reduces to the
normal 50 percent.
5) One-third salary to an officer with a non-service
connected disability who has over 10 years service. 4
The health services provided are as follows
1) Sick leave benefits provide an officer off duty for
illness 60 days sick leave without loss of pay. If his
sickness continues, he is entitled to half pay for an
additional 60 days, and if his sickness further continues,
zie receives such pay as the city council may direct.
2) Medical plans are provided for the employee and
dependents. The City pays $15 monthly, which approx-
imates the employee's contribution. Dependents are
covered at the employee's expense.
3) Dental plan coverage is paid in full by the City for the
employ,7.e only.
4) An officer off duty for injuries rcceived in the line of
duty receives whatever medical, surgical or hospital
treatment, including medicine, nursing, and medical
and surgical supplies, he may need during the period
132
of disability, remaining on a full pay status until
return to duty or disability retirement, 5
Fifteen working days per year of vacation are provided,
increasing to a maximum of 20 working days. In addition,
officers are compensated for 11 holidays per year. 6
There are also several educational incentive programs,
providing cash payments to officers completing approved college
course work or continuing their education. 7
We have no comparable estimates of the pecuniary value
of benefits available to the average San Francisco-Oakland
worker. However, the police benefits, particularly the retire-
ment provisions which enable a man who joined the force at age
21 to retire at age 46 at one-half pay, should certainly compare
favorably, adding positively to the relative pecuniary attractive-
ness of the police job.
VACANCIES
Yet, the relative pecuniary attractiveness of the job did
not appear to be sufficient to offset an apparent distaste for
police work in the population, at least in the population of
applicants of interest to the Oakland Police Department. Oakland,
along with other major cities, complained bitterly in the 1960's
of shortages of qualified applicants, of an inability to fill vacant
133
positions quickly. 8 The city was very unhappy, for example,
with vacancy rates of over 6 percent of authorized positions which
prevailed in 1966 and 1968, in spite of elaborate nationwide re-
cruitment efforts.
This unhappiness contributed, of course, to the granting by
the city of the double wage increase to patrolmen in 1968, 8.7
percent in July and 6. 5 percent in August. This increased patrol-
men's starting salaries from 117 percent of average earnings of
production workers in manufacturing to 125 percent (see Table
4-7). Simultaneously, general econc nic conditions worsened,
political turmoil lessened, departmental image improved, and
attempts to attract more minorities to the force were launched.
The medicine of increased salary, aided by other conditions,
worked. The vacancy rate declined from 6.1 percent of authorized
positions in July, 1968, to zero in June, 1969. It remains very
small, in spite of the abandonment of national recruiting and the
adoption of residence preference. The history of the vacancy
rate from 1960 to 1973 is presented in Table 4-12. The complaints
today are not of general shortages of qualified applicants, but
only of shortages of qualified minority applicants.
134
Table 4-12. Oakland police officer vacancies, 1960 to 1973
Date
Number of Number ofAuthorized Filled Number of VacancyPositions Positions Vacancies Rate
June 13 toJuly 24, 1960 543 494 47 .087
June 26 toAugust 6, 1961 543 459 84 .155
May 28 toJuly 8, 1962 488 450 38 .078
June 11 toJuly 21, 1963 468 457 11 .024
June 23, 1964 468 465 3 .006July 5, 1965 468 458 10 .021July 11, 1966 511 479 32 .063July 19, 1967 511 496 15 .029July 1, 1968 511 480 31 .061June 30, 1969 532 532 0 000June 29, 1970 525 524 1 .002July 5, 1971 525 524 1 .002July 3, 1972 537 514 23 .043July 2, 1973 535 523 12 .022
Note: The vacancy rate is computed by dividing the number ofvacancies by the number of authorized positions.
Source: Oakland Police Department, Personnel Division, memofrom Lieutenant J. McArthur.
FACT AND THEORY
The disappearance of shortages following a marked increase
in relative wages seems to cast the final blow to those theories
which sought to explain the police shortages of the 1960's solely
through a monopsony model implying job vacancies in equilibrium
135
at a monopsony wage. More likely, if a simple, model is desired,
we were observing an Arrow and Capron dynamic shortage, in
which the lags in market adjustment were ultimately corrected.
Even some of Ronald Ehrenberg's inferences must be
modified. An increase in the average wage of Oakland policemen
did not increase the expected vacancy rate, as he suggested might
be possible, but decreased it, as simple conventional theory
would predict. However, the Ehrenberg model has sufficient
complexity and scope to remain interesting. He did not argue,
after all, that vacancy rates had to increase if wages increased,
only that they might. One would have to consider also changes in
the value of the service to society, the number of applicants, the
authorized employment level, the quality distribution of appli-
cants, the minimum acceptable quality level, and the proportion
of applicants that are accepted. In fact, for example, if the wage
increase is accompanied by higher minimum standards and/or
an upward shift in the quality distribution of applicants, then the
vacancy rate will decrease. Ehrenberg concludes simply that
without numerical estimates of the magnitude of the offsetting
impacts of an increase in the average wage on the vacancy rate,
we can presume nothing about the nature or direction of their
relationship. Nothing in this discussion is meant to suggest
136
that one experience in Oakland can prove or disprove a theory.
We are only trying to assess which theories seem most con-
sistent with our observed case.
As for the relationship between wage increases and
authorized employment levels, our evidence is ambiguous. The
large wage increases between July, 1968, and July, 1969, which
increased patrolmen's starting salaries from 117 percent of
average earnings of production workers in manufacturing to 125
percent, were accompanied as well by an increase of 21
authorized positions. Between July, 1969, and July, 1970,
relative wages jumped again, to 135 percent of production
workers' earnings, and authorized positions increased by 12.
In July, 1972, and July, 1973, both relative wages and authorized
positions remained stable, but salary increases reflecting in-
flation continued. In July, 1974, in spite of continuing stability
of the relationship of police salaries to production workers'
wages, 20 patrolman positions were cut, reducing authorized
positions to 515. This could well be an employment response to
continued salary increases brought about by the pressures of
inflation. The magnitude of the cut suggests strongly that the
response of public employment to wage increases may be stronger
than people think.
137
HIRING STAN DA R DS
We turn now to a description of the vector of hiring
standards that constitutes the minimum acceptable quality level
of the Oakland Police Department. It is these standards ti.lt
filter and restrict the available labor supply. It must be stressed
here that even in the 1960's there was no shortage of applicants
to the Oakland Police Department, only a shortage of applicants
deemed qualified. Table 4-13 shows the percent of all applicants
actually appointed to the department between 1960 and 1970.
Table 4-13. Percent of all applicants appointed to the OaklandPolice Department, 1960 to 1970
Year Percent Appointed
1960 0.761961 2.68196 2 3.811963 2.751964 4.301965 6.341966 8.531967 3.781968 5.951969 6.301970 3.79
Source: Oakland Civil Service Office
Clearly, selection is at work.
Logically, selection standards for a job should be derived
138
irorn a Clear e%-aluation of the qualities nocessary for successful
job performance. A typist must know how to type; an acceptable
standard in terms of words per minute and accuracy can be
defined. However, as we discussed at length in Chapter a, the
policeman's job is more loosely defined and often involv.es
elusive and contradictory skills. Moreover, we are not quite
sure what we mean-by successful job performance.
In spite, even perhaps because of the difficulty in specifying
desirable traits, elaborate, detailed hiring standards evolved for
policemen. The standards conformed to the images of policemen,
police administrators and civil service personnel of the good
policeman. They translated these images into decision rules
which then were reinforced by widespread usage, advocacy, and
belief in their effectiveness. They were hallowed by tradition,
but not by empirical validation of their relationship to job
performance.
The selection standards and hiring procedures of the
Oakland Police Department were quite representative of those in
police departments across the country in the 1950's and 1960's.
We will enumerate and explain them as best we can.
First, there are the requirements explicitly listed on the
job announcement posted by the city. A potential applicant
139
should be able to ascertain himself whether or not he meets
these requirements. If he knows he does not, he will probably
not apply. If he is uncertain and submits his application, the
city should be able to ascertain quickly whether the require-
ments are 'satisfied or not. They are:
1) Age: 21 to 32, to age 35 with 2 years of college or
2 years of police experience.
2) Height: 5'7" to 6'7" measured without shoes.
3) Weight: 135 lbs. , must be proportional to height.
4) Vision: 20/70 uncorrected, correctible to 20/20 one
eye, 20/30 the other, normal color vision.
5) Education: high school graduate or GED with a total
score of 262.
6) Driver's license: possession of a valid unrestricted
license by date of hire.
Age
Minimum age requirements of 21 are widely adopted by
police departments because this is the age at which a person's
full legal rights and obligations as an adult United States citizen
commence. Maximum age requirements are widely justified as
necessary for the financial solvency of police retirement systems.
Clearly at work here as well are American mystiques about young
men in the prime of physical fitness.
140
All of California's 288 distinct law enforcement agencies
have minimum age requirements: all but 7 have maximum age
requirements as well. But the details of these requirements vary
widely. Mandated minimum ages rang,,. from 20 to 25, mandated
maximums from 25 to no maximum. 9 The most common age
standards among California law enforce'ment agencies are
enumerated in Table 4-14. We list, in order of popularity, only
those requirements chosen by more than five agencies.
Table 4-14. Common age requirements of California lawenforcement agencies in 1968
Age Range Number of Departments
21 to 35 13321 to 34 26
21 to 40 2421 to 31 19
21 to 45 9
21 to 32 8
21 to 3321 to no maximum 7
21 to 39 6
21 to 30 5
other 44
Total 288
Source: State of California, Commission on Peace OfficersStandards and Training, A Career in Law Enforcement,October, 1969.
141
Height and Weight
Height and weight requirements also come from American
popular beliefs about what constitutes the physically fit male.
The minimum requirements satisfy the stereotype of the large
male police officer, called upon to perform duties in which size
and presence are assets. It is argued, for example, that size
may be needed to restrain hostile or struggling persons during
arrest or flight, to lift or carry persons, to pull persons from
wrecked automobiles, to subdue suspects, to climb fences, to
kick in doors, etc. 10 More sophisticated arguments are made
as well. One is that physical presence alone may enable a
policeman to dominate a situation that he might not otherwise be
able to control. Another is that small officers are more likely
to be involved in physical altercations because they present less
imposing physical prowess. Similarly, they may be more often
called upon to use lethal methods of self defense. 11 These
beliefs are deeply imbedded and passionately held in police
circles, despite the lack of empirical verification of their
reality.
The range of standards, however, alleged to satisfy these
needs is wide. Over 100 different combinations of height and
weight minima and maxima are embraced by the 288 California
142
Table Common height and weight requirements ofCalifornia law enforcement agencies in 1968
Height Rar:ge Weight Range Number of Departments
5'9" to proportional 27
5'9" to 150 lbs. minimum 19
5'8" to proportional 18
proportion tl proportional 145'8" to 6'6" proportional 13
5'8" to 6'5" proportional 11
5'9" to 6'6" proportional 9
5'8" to 6'4" proportional 8
5'8" to 1505'8" to 145 8
5'9" to 6'5" proportional 7
5'9" to 6'4" proportional 7
5'8" to 6'6" 150 6
5'9" to 6'6" 150 5
other combination 128
8
Total 288
Source: State of California, Commission on Peace OfficersStandards and Training, A Career in Law Enforcement,Dctober, 1969.
law enforcement agencies. We list those chosen by more than
five departments, again in order of popularity, in Table 4-15.
Vision
Visual acuity standards conform to what is believed to
be necessary to perceive and describe situations and evidence
accurately, to defend oneself in possible combat, and to be an
expert marksman under all conditions. Departments became
143
less insistent on virtually perfect uncorrected vision after
contact lenses, believed more reliable than glasses, were
perfected. Table 4-16 presents the most popular visual acuity
standards in California.
Table 4-16. Common visual acuity standards of Californialaw enforcement agencies in 1968
Visual AcuityUncorrected Corrected
Vision to Number of Departments
20/40 20/ 20 120
20/70 20/30 53
20/30 20/20 25unspecified 20/ 20 23
no standard 18
20/30 unspecified 11
20/70 20/ 20 7
other standard 31
Total 288
Source: State of California, Commission on Peace OfficersStandards and Training, A Career in Law Enforcement,October, 1969.
Education
All California police agencies require either a high school
degree or a passing score on the Army's high school equivalency
examination, the GED (test of General Educational Development).
Oakland's requirements here exceed the minimum only by
144
specifying an acceptable minimum passing score on the GED,
262, which means the candidate scored higher than 70 percent of
the people taking the test. It is taken for granted that high school
graduation represents a minimum achievement necessary for
acceptable trainability, performance and character. More
controversial is whether or not the standard should be raised to
require some college education.
In favor of raising educational qualifications is the argu-
ment that college experience provides essential understanding of
the complexities of the economic and social environment in which
policemen function, and that, therefore, more educated policemen
will be more professional, exercise discretion more intelligently,
and perform better generally. Moreover, policemen should not
fall behind the rising educational level of the population as a
whole. On the other side are the arguments that higher educa-
tional levels would exclude many young men who would make fine
policemen, that police departments, through their own recruit
schools, can provide the necessary general knowledge and
specific training, that on-the-job experience is the best teacher,
and that the relationship between education and job performance
is ambiguous. Also, policemen can be encouraged, through
educational incentive programs of all kinds, to continue their
education while on the force.
145
Less than 10 percent of California police agencies require
more than high school graduation. Of the 28 agencies that have
higher standards, three are State of California agencies, nine
require less than 30 units, eight require 30 to 59 units, six
require 60 units, one demands an A. A. degree and one demands
a B. A, degree. 12 The departments with higher standards
are of most interest to Oakland are Berkeley, San Jose, and
Walnut Creek, all of which ask for 60 units of college.
However, as we will show in detail in Chapter 5, the
probability of successfully passing through the rest of the Oakland
Police Department's testing process, particularly the Mental
Adaptability test, is increased by large magnitudes as years of
higher education increase from some years of junior college
through three to four years of senior college or a college degree.
So minimum requirements do not quite tell the whole story.
Filing the Employment AEplication
If a potential applicant feels that he meets all the specified
requirements, he files an employment application. He is asked
his name, address, age, sex, marital status, height, weight,
ti firth date, citizenship status, number of dependents, and
residence. He is asked if he was ever discharged or forced to
resign because of unsatisfactory service. He is asked if he has
146
ever been "arrested, indicted, summoned into court, convicted,
imprisoned, placed on probation or required to pay a fine of more
than $25." 13 He is asked for a complete military history,
educational history, employment history and medical history.
The application is then reviewed by the Civil Service Office staff
and either approved or disapproved. It may be disapproved for
a number of reasons. First, it may be clear that the applicant
does not meet a stated requirement, for example, he is over-age,
overweight, under-educated, etc. Secondly, the application may
indicate an undesirable citizenship history, which, upon further
investigation, is judged unacceptable. Reasons for rejection
here include the nature of the arrest record, excessive traffic
violations, avoidance of alimony or child support payments, or
less than honorable discharge from the military. Thirdly, the
application may indicate an undesirable employment history, such
as excessive job changing, or discharge for cause. Finally,
certain medical conditions will immediately disqualify a candi-
date, such as a history of back or knee injuries, epileptic attacks,
asthma, nervous disorders, etc.
Acceptance of the application means that an individual will
be allowed to proceed further through the testing process. If he
continues to succeed, all his application statements will be
147
reviewed again in greater depth in the background investigation
and medical examination at the end, Lack of truthfulness here
will always be uncovered, and will probably, in itself, he cause
for disqualification. A rejected applicant may try to clear up
his difficulties, if possible, and apply again.
Approximately 20 percent of all applications are rejected.
This means the applicant can proceed no further. Table 4-17
shows the number of applications filed and the number accepted
from 1960 to 1970.
Table 4-17. Number of employment applications for Oaklandpatrolmen filed and accepted, 1960 to 1970
DateNumber
FiledNumber
AcceptedPercentAccepted
1960 1, 185 1, 049 89
1961 934 787 84
1962 1, -...)6 963 80
1963 :8C0 726 91
1964 c 9 9 809 81
1965 ..'78 837 86
1966 961 732 76
1967 '321 549 67
1968 1, 518 1, 252 8Z
1969 1, 415 1, 093 771970 1, 609 1, 149 71
Source: Oakland Civil Service Office
148
Mental Adaptability Test
Once the employment application has been accepted, the
applicant progresses to the Mental Adaptability test. Until
November, 1970, this test was either the Otis Self-Administering
Test of Mental Ability or the Army General Classification test,
both general multiple choice tests of verbal, mathematical and
reasoning aptitude. The Otis test, for example, was copy-
righted in 1922 and standardized in 19'i2 in Oakland. Items were
selected at that time to discriminate between a good group and a
poor group of students. Both groups had reached the same
average educational status, but the good group was two years
younger than the poor group. "Each item justified its inclusion,
therefore, because it distinguished between students who pro-
gressed slowly and those who progressed rapidly. "14
A candidate had to achieve a score of 52, equivalent to
an IQ of 110, on the Otis to pass. The passing point was chosen
to indicate people who are successful with college curriculums.
The passing point is justified considering the great amounts of
material we expect a patrolman to learn and understand. "15
In November, 1970, the old and allegedly culturally-biased
Otis test was replaced by a new, less culturally-biased version,
the Otis-Lennon Mental Ability Test, emphasizing "measuring
149
the pupil's facility in reasoning and in dealing abstractly with
verbal, symbolic, and figural test content sampling a broad
range of cognitive abilities. "16 .The standardization procedures
for the new test "were designed to yield a stratified random-
cluster sample of pupils representative of United States school
pupils enrolled in grades K to la. The controls used in the
selection of these school systems were designed to provide one
of the most representative norming studies ever undertaken for
an entire test series of this type. Variables used in stratifying
school systems were: 1) enrollment, 2) type, i, e. , public,
private, church-related, 3) a composite socioeconomic index
based upon median family income and median years of school
completed by adults 25 years and older as reported for the
census unit served by the school system, and 4) geographic
Iregion.
7 The test items were changed, but the format was
similar to the old Otis, except for a greater proportion of non-
verbal or figural items.
A candidate had to achieve a score of 45, representative
of the midpoint or fiftieth percentile of the nation, equivalent to
an IQ of 98, to pass.
Chapter 5 discusses in detail the characteristics for which
the Mental Adaptability tests actually seemed to select. We will
150
Table 4-18. Number of Oakland patrolman applicants takingand passing the Mental Adaptability Test, 1960 to 1970
DateNumber
ExaminedNumberPassed
PercentPassed
1960 772 487 631961 506 205 411962 564 315 56
1963 534 179 341964 629 283 451965 606 268 441966 592 297 50
1967 439 266 61
1968 765 360 471969 56.6 351 621970 595 327 55
Source: Oakland Civil Service Office
defer further discussion until then, except to indicate that only
about half the candidates taking this test passed. The number
taking the test and the number and percent passing between 1960
and 1970 are presented in Table 4-18.
Civil Service Test
The applicant who has not been weeded out moves on to the
Civil Service Test. This is a test written by the Civil Service
staff to measure general knowledge. It was designed to "measure
a candidate's knowledge on a wide variety of subjects since it was
assumed that the candidate who is well informed will make a
151
better policeman. "18 The typical test contains 100 multiple
choice questions, 50 on general knowledge, 35 on current events,
and 15 on analytical ability, usually reading comprehension. A
score of 70 out of the 100 questions was needed to pass.
About 82 percent of the candidates who took the test passed
it. This test, too, will be analyzed in detail in Chapter 5.
Physical Agility Test
The Physical Agility Test is revealed on the job announce-
ment for the patrolman examination which is posted around the
city and enumerates the other requirements. It consists of five
exercise events to be completed in order with approximately three
minutes rest between events. Candidates are scored on form as
well as number completed. The events are:
Squat thrusts: 25 required in one minute's time
Sit-ups: 25 required
Push-ups: 22 required
Squat jumps: 27 required
Pull-ups: 6 required
The Physical Agility Test, like the height and weight
requirements, rests on the assumption that strength and physical
skills may be needed on the job, and candidates who are insuffi-
ciently fit should be dropped. Between 1960 and 1970, 86 percent-
152
of the candidates who took the Physical Agility Test passed
it.
Oral Examination
Next, the candidate is scheduled for an oral examination,
which only 54 percent passed between 1960 and 1970. The inter-
view is conducted by two police officers and one representative
of the Civil Service Office. The candidate is judged on appear-
ance, bearing, and manner, ability to present ideas, social
adaptability, alertness, judgment, emotional stability and
interest. If he fails in any one area, he is judged clearly un-
acceptable. If he gets an overall rating of less than 70 percent,
he does not pass.
The interviewers have the candidate's employment applica-
tion in front of them and probe his personal history in depth. He
will be asked about any problems suggested by the application,
such as a divorce or separation, a complicated work or school
history, reasons for job changes, etc. The honesty and attitudes
revealed by his approach to the questions are as important as
the answers themselves.
In the early 1960's, a great deal of emphasis was placed
by oral boards on appearance and bearing, on command presence,
on how impressive a man might look in uniform. Stress inter-
153
view techniques were also used. The caLididate was placed in
a hypothetical difficult situation and asked to respond to it; the
interviewers would then seize his answers and try to box him
into an even more difficult situation. They were seeking to
evaluate his logical thought processes when making decisions,
his moral fiber, and how well he "stood up to the heat."
Stress interview techniques and emphasis on command
presence have been dropped in favor of attempting to find out
what the candidate know s about Oakland and the police job.
Attitudes toward the minority community and its problems are
probed. The board watts to know why the candidate is applying
for the job, what other jobs he has applied for, why he is
interested in Oakland. It seeks to evaluate, as well, his
reasonableness. Can he accept, for example, that there are
things in the world that he cannot handle, that he is not going to
save the world? The oral also serves to screen out the clearly
maladjusted or psychologically disturbed.
We will have some more to say about the results of the
oral in Chapter 5.
Medical Test
After passing the oral, a candidate is scheduled for an
extensive medical and laboratory examination by the city
154
physician. Twenty percent of the candidates are eliminated by
the doctor, whose decisions are final. The doctor evaluates the
candidate's physical condition both from the point of view of his
ability to perform a strenuous and stressful job successfully and
of the city's financial interests. Because of the extensive health
benefits enjoyed by policemen, the city is very careful to
minimize its risks and to refuse employment to anyone with a
potential medical problem. All doubts here are resolved by the
physician in favor of the city.
Psychiatric Evaluation
The psychiatric evaluation consists of clinical paper-and-
pencil tests, a Rorschach test, and a personal interview by the
department's psychiatrist, which discusses such questions as
the candidate's most memorable dream. It is most effective in
revealing any extreme personality deviance that was not picked
up in the oral. Designing good psychiatric tests is a subject all
its own, which we will not begin to tackle here. The experience
of the Oakland department psychiatrist has led him to believe
that some potentially effective indicators of successful police
performance are a non-defensive Rorschach and intense affect
reported in dreams; indicative of ability to feel. 20 Also positive,
he thinks, are such background factors as being a family man,
155
having a stable childhood with no step-father and an active
father, and having a positive social attitude. 21 In any case, 96
percent of the candidates pass the psychiatric evaluation.
Background Investi ation
The last hurdle is the background investigation. An Oak-
land police officer is assigned to check out the candidate thorough-
ly. He visits the candidate's neighborhood and makes inquiries
of friends, neighbors, local businessmen; he interviews the
candidate's wife and family; he visits schools attended and past
employers. Criminal and credit records are checked. All leads
are followed. Eighty-eight percent of the candidates between
1960 and 1970 passed the background investigation. All background
rejections must be reviewed and approved by the chief of police.
Rejected candidates have the right to appeal.
Barriers to Entry
Clearly, all these hiring standards and tests combine to
restrict seriously the potential available labor supply. The
Oakland Police Department itself did some rough calculations
of their effects. The conclusion was that less than half of one
percent of Oakland's population in 1969 would be eligible to
commence the testing and approximately one-tenth of one
percent of Oakland's population would pass successfully through
156
the selection process. If the male population 21 to 32 years of
age is used as a base, still only slightly over 2 percent of that
group would succeed in the tests. 22 It is no wonder that
national recruiting campaigns were launched.
We turn now to the empirical analysis. We shall describe
the economic characteristics of the applicants who are success-
ful on the Mental Adaptability, the Civil Service Test, and the
oral examination.
157
CHAPTER 4
FOOTNOTES
1. U. S. Department of Commerce, Social and EconomicStatistics Administration, Bureau of the Census, 1970 Censusof Population, Detailed Characteristics, California, Section1 (PC (1)-D6Calif.), pp. Calif. 6-1823 to 6-1828, Table 175.
2. U. S. Dennrtment of Commerce, Social and EconomicStatistics Administration, Bureau of the Census, 1970 Censusof Population, Detailed Characteristics, California, Section 2(PC (1)-D6Calif.), pp. Calif. 6-2336 to 6-2338, Table 197! and1970 Census of Population, General Social and EconomicCharacteristics, California, Appendix B, pp, app. 25 to app.27.
3. Estimates provided by the Oakland Civil Service Office
4 Taken from the City of Oakland announcement for PatrolmanExamination, September, 1970.
5. Ibid.
6. Ibid.
7. Ibid.
8. See Raymond L. Bancroft, "Municipal Law Enforcement,1966," Nation's Cities, February,1966, pp. 15-17. "TheNation's police departments are encountering seriousdifficulty in maintaining their forces at authorized strength.A survey conducted by the National League of Cities in 1966disclosed that over 65 percent of the departments surveyedwere understaffed; that these departments were 5, 840officers, or 5 percent, below authorized strength; and were11, 864 officers, or 10 percent, below preferred strength...Since, on the average, police departments are currently 5percent below authorized strength, since the authorizedstrength of police departments has increased at the rate ofapproximately 3 percent each year, and since an average of5. 4 percent of existing personnel leave their departments
158
each year (due to resignation, dismissal, retirement, ordeath), 50, 000 new police officers will be needed in 1967alone. "
9. State of California, Commission on Peace OfficersStandards and Training, A Career in Law Enforcement,October, 1969.
10. See the arguments in defense of height and weight require-ments in the recent local court cases: Veragene Hardy vs.William F. Stumpf, 1 Civ 32689, Court of Appeal, State ofCalifornia, First Appellate District, Division 3., March 15,1974; and Joan Hail vs. James S. White, U. S. DistrictCourt, Northern California, September 19, 1973. See alsoJ. Stanley Pottinger, Brief for the United States as AmicusCurie, No. 74-1038, filed with U. S. Court of Appeals, NinthCircuit, April 5, 1974, in the pending appeal of Hail vs.White, 29 pp.
11. See the testimony of Oakland Police Sergeant Anderson andInternational Association of Chiefs of Police researchscientist Terry Eisenberg in Hail vs. White, op. cit.
12. State of California, Commission on Peace OfficersStandards and Training, A Career in Law Enforcement,October, 1969.
13. City of Oakland, Employment Application.
14. Arthur S. Otis, Otis Self-Administering Tests of MentalAbility, Manual of Directions and Key (Revised) for Inter-mediate and Higher Examinations, New York: Harcourt,Brace and World, 1922.
15, Oakland Civil Service Office memorandum, "Oakland PoliceDepartment Recruitment and Patrolman Selection Process, "
December, 1964.
16. Arthur S. Ctis and Roger T. Lennon, Otis-Lennon MentalAbility Test, Manual for Administration, New York: Har-court, Brace and World, 1967, p. 4.
17. Ibid. , p. L3.
159
18. Oakland Civil Service Office memorandum, "OaklandPolice Department Recruitment and Patrolman SelectionProcess, " December, 1964.
19. City of Oakland announcement for Patrolman Examination,September, 1970,
O. Dr. Floyd 0. Due, "History and Personality CharacteristicsFavoring Success as a Police Officer," 1972 unpublishedpaper.
Zl. Ibid.
LL. Oakland Police Department, "Demographic Analysis,"unpublished memorandum.
1'60
CHAPTER 5
THE EMPIRICAL ANALYSIS OF SELECTION TESTS
This chapter attempts to specify the characteristics for
which the Oakland Police Department's selection process between
1965 and 1971 is testing. It is a beginning in observing and
defining the elusive vector of hiring standards.
THE DATA
This study is based on information gathered from the
employment applications of all applicants to the Oakland Police
Department from 1965 to June, 1971. Only individuals for whom
all relevant information was available were included. Therefore,
the total number of applicants included in the study, 2,36' is
less than the total actually tested. For example, all applicants
for whom military service represented the only employment
experience would be eliminated because no meaningful salary
observation would b,e available for them, given the incomparability
of military and civilian wages. Sufficient applicants for whom all
information is available remain so there should be no serious
inherent bias in the study arising from necessary elimination.
One critical variable is missing for the bulk of the study.
We have no indication of minority status, because the Fair
161
Employment Practices laws in this time period prohibited any
indication of race on employment applications or processes.
Today, these laws insist on the collection and reporting of such
data. Explicit data on minority status became available only in
1971. Therefore, a separate analysis of 213 applicants in the
first part of 1971 will be made in order to pay particular
attention to the minority data.
CENTRAL QUESTIONS
What characteristics would be expected to determine
success or failure on the battery of selection tests used by the
Oakland Police Department? Why would each characteristic be
included? What independent effect would it be expected to have?
How could it be measured?
We will be primarily concerned with explaining performance
on the Mental Adaptability test, the Civil Service test, and the
oral personal interview. We will consider both the determinants
of the probability of success on each test and the determinants of
the raw score. For each test, the following specific questions
will be asked: 1) passing the test is a function of what significant
variables? 2) a unit increase (decrease) in an independent variable
will raise (lower) the probability of passing the test by how much?
3) test score is a function of what significant variables? 4) a
162
unit increase (decrease) in a particular variable will raise (lower)
the test score by how much? We will consider the same set of
independent variables for each test.
DEPENDFNT VARIABLES
The dependent variables, therefore, are:
MA1 probability of passing Mental Adaptability test
MAL score on Mental Adaptability test
CS probability of passing Civil Service test
CST2
score on Civil Service test
OR1 probability of passing Oral exam
OR score on Oral exam
We will also consider briefly the probability of passimg the
Physical Agility tust, PA.
INDEPENDFNT VARIABLFS
Years of Fducation
We would expect that the probability of success on the
tests would be greater for applicants with more education, i. c.,
someone with three to four years of college would be more likely
to succeed than would a high school graduate. People with
college experience are likely to be better test performers (or
have higher measured I.Q.'s) than people without. Knowledge or
163
ability absorbed in college may help to pass tests. The debate in
professional police circles about the desirability of requiring
college education, and how much, if any, should be required,
has not been resolved. Yet the belief that better educated people
may be better able to deal well with the myriad challenges of
police work is widespread, although it has been neither proved
nor disproved. It is accompanied by the concern of police depart-
ments with professional image. Since education is a prerequisite
to professionalism, better educated men are desirable to upgrade
the department's image. Police departments also provide a
considerable amount of on-the-job training (Oakland has a 20-
week recruit school) and educational experience indicates train-
ability. (It is known from human capital literature that the better
educated receive the most on-the-job training.) Also, the depart-
ment can benefit from the general human capital already embodied
in education. Therefore, it is hypothesized that both qualities
inherent in the educated and policy decisions of police departments
which are consciously or unconsciously incorporated into selection
test design combine to increase the probability of passing as
education increases. In order to determine just how sensitive the
probability of passing is to differences in educational levels, and
because of the policy interest of the results, education will be
164
measured in rather line breakdowns. Junior college and college
will be considered separately because their products may be
quite different. The education variable is specified as follows:
J no junior college or college
J1
, less than on( year of junior college
J2 = one year of junior college
J3 more than one to two years of junior college
J4 = more than two years of junior college
J5 = A, A. degree
C0
= no college
Cl = less than one year of college
C2 = one but less than two years of college
C3 two but less than two years of college
C4 three to four years of college
C5
B. A. or B. S. degree
If an applicant reported both junior college and college experience,
the junior college years were combined with the college years and
reported in the college category. Overlapping was therefore
avoided and the junior college category includes only individuals
whose sole experience is with junior college.
Residence
The thilosophy of residence requirements is also in flux.
165
Historically, rules requiring policemen and other city employees
to live in their city were common. The movement of the middle
class to suburbia and the rise of the concept of the professional
policeman led to an elimination of residency requirements.
Oakland, in the forefront of the trend, dropped its residency re-
quirement in November, 1955, and launched a campaign of nation-
wide recruitment of police which became increasingly extensive
until 1969. Today, the pressures of unemployment and minority
politics are leading some cities to consider re-adopting residence
rules for city employment. Oakland, in fact, added a residence
preference rule in the fall of 1971. It dropped its elaborate
nation-wide recruiting campaigns in 1969 in response to the
pressures mentioned above and to anti-police demonstrations
encountered on some campuses.
Residence at the time of application will be classified into
three groups:
R0
out-of-state residence
R1 , Oakland area residence
R2 other California residence
On one hand, Oakland area residents might be expected to
have a higher probability of success than others because of a
greater degree of f am i 1 iarity with Oakland and its problems and
166
police force. On the other hand, people from out of the area have
invested considerably more time in search activity to discover the
Oakland Police Department and become interested. Hence, they
may be more committed, motivated or professional than an
Oakland resident who needed only to come in off the street. Also,
people from out of the area are more likely to have assessed their
chances of success before applying; the higher cost of applying
for them would lead to more self-selection. It seems unclear
a priori whether the positive or negative effect of living in the
Oakland area will prevail.
Minority Status
It is well known that members of minority groups are less
likely to pass standardized written tests which are not "culture
free" than majority members, at the same passing point. There-
fore, we would expect, other things being equal, minority status
per se would lower the probability of passing the tests. Unfor-
tunately, except for the first six months of 1971, we do not have
data to test directly this critical question. We will analyze the
first few months of 1971 separately to cast the light of even a
small amount of data on the matter. For that analysis
MD 7, majority member
MI black, Spanish-speaking, Oriental, Filipino,
and other minority.
167
Whereas it would he desirable to analyze each minority group
separately, there is an insufficient amount of data to attempt to
do so.
In an attempt to analyze the influence of minority status
even indirectly, proxies were sought for the study as a whole.
It is hoped that given the pattern of migration of blacks to
Oakland, a disproportionate number of applicants who attended
high school in the South would be black, and theref ore a variable
indicating region of high school attendance would pick up some of
the influence of race. However, we also know that education in
the South may be inferior even for whites, so the variable may
be measuring only generally poor schools in the South. In either
case, attending high school in the South would lower the probability
of passing the test. The variable is probably measuring some
combination of race and inferiority of schools.
H0
= attended high school in West, North Central or
Northeast
H1
attended high school in South
Specific Police Training or Experience
It could be hypothesized that specific police training or
experience as manifested by studying police science in college
or attending non-college police school or previous employment
168
as a policeman would increase an applicant's probability of
success. The department could benefit from the specific human
capital embodied in the person. Moreover, motivation, interest,
sincerity, would be demonstrated by the individual's own invelot,
ment in acquiring specific training. The effect of studying police
science in college or attending non-college police school is
expected to be weaker than that of actual ex-perience as a police-
man. Since the department has its own elaborate training program,
the courses may be extraneous and bear the risk of having to
reverse misconceptions. Experience as a policeman, however,
implies at least some knowledge of the realities of the job,
although perhaps in a very different setting, some concrete ex-
perience, and an application choice based on rational assessment.
Also, the man is likely to have passed a similar body of tests
elsewhere and the department can reap the benefit of another
department's appraisal and performance evaluation.
PS0 = did not study police science in college
PS1 = studied police science in college
PSC° = did not 'study in non-college police school
PSC, = studied in non-college police school
P0
= no previous employment experience as a policeman
P1 previous employment experience as a policeman
169
Previous Wagt.
It is hypothesized that applicants from higher wage groups
will be more likely to succeed than applicants from lower wage
groups, other things equal. Past wage or salary classification
should be indicative of alternative worth, human capital accumula-
tion, or marginal productivity. Other things equal, an individual
from a higher salary group should be more valuable and desirable
to the police department than an individual from a lower salary
group. Also, higher salaries indicate past success and possession
of such all-American virtues as ambition, motivation, productivity,
etc. The interpretation here obviously has to be modified if some
of flake observations on the lower end of the scale represent people
working part-time white in college or other non-standard depar-
tures from full-time entry into the wort of work.
W0
--- earning less than $300 per month on last job
W1
$300-$499 per month on last job
W $500-$699 per month on last job
W3 $700 or above per month on last job
We have not corrected for the general rise in wages between
1965 and 1971, We assume our broad categories retain their
validity.
170
Number of Jobs Fnamerated
More jobs enumerated indicate more experience which may
be desirable and increase the probability of success. Alternatively,
except for such jobs as carpenter, it may indicate instability or
emotional problems, and therefore decrease the probability of
success, particularly on the oral exam. It is not clear a priori
which effect will predominate. Number of jobs is included here
as a continuous variable.
E = number of jobs enumerated
Age
The older a person is at a given educational level, the less
likely he is to pass a standardized test. Knowledge and test-
taking ability depreciate. Therefore, we expect age, entered as
a continuous variable, to have a negative effect on the probability
of passing the first two tests. However, if the oral is testing for
maturity, among other things, age may have a positive effect on
the probability of passing the oral.
A = age (in years)
Physical Qualities
Physical qualities may have some effect on ability to pass
the physical agility, which will be discussed briefly, so we will
include:
171
B = height (in inches)
C = weight (in pounds)
METHODOLOGY
The effect of a unit increase (decrease) in an independent
variable on the probability of passing each test and on the raw
score will be explored through a multiple regression analysis.
For the probability equations the dependent variables will take
on the binary values 0 = failure, 1 = success. For the raw score
equations the dependent variables will have the conventional
continuous form. The problem with using regression analysis
for the dichotomous case is that the Gauss-Markov assumption
that V = I, i, e. , that the disturbance variances are constant from
observation to observation (homoscedastic) is untenable. It is
also possible, since the model is linear, to obtain probabilities
greater than one. A probit analysis model would correct these
problems. A probit program was obtained and experimented
with, but unfortunately proved to be inordinately expensive, at
least three times the cost of the multiple regression program.
Moreover, for those equations run on both programs, the
probit and regression analyses yielded the same results. Some
of the equations reported for the small sample separate analysis
172
of 1971 data do have the problem of probabilities greater than
one. However,. only multiple regression results will be reported.
The equations will contain two continuous variables, age
(A) and number of jobs enumerated (E). All other independent
variables will enter as sets of dummy variables as defined above
and sumnlarized here for convenience:
, no junior college or college
J1 = less than one year of junior college
J2
= one year of junior college
J3 = more than one to two years of junior college,
J4 = more than two years of junior college
5= A. A. degree
C0 = no college
C1 less than one year of college
C2 ont but less than two years of college
C3 = two but less than three years of college
C4 = three to four years of college
C5 = B. A. or B. S. degree
0= out-of-state residence
R1 = Oakland area residence
R2 -- other California residence
MO= majority member
173
M1 = minority member
HO = attended high school in West, NortbNCentral or
11Northwest
H1
= attended high school in South
PS0
= did not study police science in college
PS1 = studied police science in college
PSC0
= did not study in non-college police school
PSC. = studied in non-college police school
0= no previous employment experience as a
policeman
P1 = previous employment experience as a policeman
W0 = earning less than $300 per month on last job
W1 = $300-$499 per month on last job
W = $500-$699 per month on last job
W3 = $700 or above per month on last job
number of jobs enumerated
A = age
K = constant
In each case, the X0 category (J0, Co, R0, H0, PS0, PSC°,
P0, W0) indicates the reference category which is omitted from
the regression in order to avoid a perfect linear relationship
among the dummy variables in each set.
174
The equations, therefore, take the following form:
c 2C..2
+ c3
C3 + c4 C4 + c5 _c5 +d1R + da 112 e 1H1
11 PS1
+ g IPSCI h IP 1i1W1 i2 w2 + 13 W3
where b1 b5 represent the net (partial) regression co-
efficient
c1 c5
of dummy variables Ji J5, C1 C5, etc.
etc.
Each regression coefficient is to be interpreted as an
estimate of the net effect on the probability of an applicant being
successful of his belonging to the particular category in question
rather than to the reference category, all other variables remaining
constant. For example, a coefficient of .33 on C4 would mean that
having three to four years of college would increase an applicant's
chance of passing the MA by .33 probability points over his chance
of passing if he had no college, ether things equal. For example,
suppose that if an applicant were in the reference group in all
categories, he would have a . 50 probability of passing; if he
remained in the reference group in all categories but education,
and now had three to four years of college, his probability of
passing would now be .83, surely a significant increase.
The raw score equations (MA2, CST2, OR2) are of
175
identical form but deal with actual raw scores. A coefficient of
9. 71 on C4 would mean that, other things equal, having three to
four years of college rather than no college would raise the score
on the MA 9.71 points, say from a reference group mean of 60
to 69. 71.
A word of caution should be added in interpreting the raw
score equations. Results will be reported based on the experience
of all applicants from 1965 to 1971. The meaning of the score
does not remain the same, but changes throughout the time period,
particularly between test revision dates. As the test ages, items
lose validity and may be eliminated and the raw score passing
point may be lowered from, say, 55 when the test was designed to
48 eight months later. This is a problem particularly for the
CST. Also, for the purpose of research and to determine if
lowering pass points on test scores lowered the quality of man-
power, the minimum passing point on the MA was lowered from
52 to 48 to 40, during the time period reported. These problems
could be dealt with by analyzing each relevant period separately,
but the results reported here will be aggregate and incorporate
these difficulties.
If the probability equation results and the score equation
results, each with their own technical difficulties, seem to
176
indicate the same conclusions, perhaps the technical difficulties
are not critical for our purposes.
RESULTS
Mental Adaztability
The results for the Mental Adaptability test are based on
the information contained in the employment applications of 2, 36b
individuals tested between 1965 and 1971. Table 5-1 summarizes
the number applying in each category, the number successful in
each category, and the percent successful in each category.
Of those who did not attend junior college or college,
25.7 percent were successful. We can see, therefore, by looking
simply at the percent distribution that education makes a differ-
ence; 87.6 percent of those with a B.A. or B.S. degree and
83.9 percent of those with three to four years of college passed
the test. Percentages are useful descriptive devices. Multiple
regression analysis will enable us to say how much of the
increased probability of passing the MA is due to the independent
effect of having three to four years of college or to having a B.A.
or B.S. rather than no college, holding all other variables
constant.
The multiple regression results for the probabilities of
177
Table 5-1. Percent of applicants in each non-reference categorysuccessful on the Mental Adaptability
Category
TotalNumber
Applicants
TotalNumber
Successful
Percent ofApplicantsSuccessful
Total 2,366 1,375 58.1J1 422 251 59.5Jz 14g 95 64.2J3 232 154 66.434 244 187 76.6J5 44 31 70.5Cl 156 112 71.8C2 110 90 81.8C3 38 29 76.3C4 87 73 83.9C5 194 170 87.6R1 1,315 710 54.0R2 713 447 62.7H1 302 109 36.1PSC1 91 59 64.8PSI 437 295 67. 5
PI 213 151 70.9W1 917 506 55.2Wz 754 423 56.1W3 256 178 69.5
passing each test and the raw scores on each test are summarized
in Table 5-2 on the following page. The results for the probability
of an applicant passing the Mental Adaptability test are found in
column 1 of that table. Each coefficient represents the increase
or decrease in the probability of passing resulting from being in
the specified category, rather than in the reference group. For
example, the probability of a person in category C4 passing the
178
Table 5-2. Regression coefficients, all applicants, 1965 toTune 1971
(1)MAI
(2) (3)CST 1
1 1
(4)MA2
(5)CST 2
(6)OR2
K . 799 * . 811- . 507**(. 094) (. U /1) (. 145)
A . 015** . 002 . 008(. 004) (. 003) (. 006)
JI
. 083** . 013 . 061(.027) (.021) (. 043)
J .121** .041 . 055(.041) (.030) (. 060)
. 124** . 049 . 023(.036) (. 026) (. 052).134 ** . 064** -. 053
(.036) (. 024) (. 048). 116 . 015 . 097
(.073) (.051) (. 111).180 ** .045 .041
(.039) (.027) (.054).294 ** . 038 . 060
(. 046) (. 030) (. 061). 252'; . 071 . 035
(.076) (. 051) (. 108).334V* . 102** . 143*
(. 052) (. 033) (. 069). 347 1/4* . 081** . 008
(.037) (.024) (.049)-.025 .022 -. 124*(. 031) (. 023) (. 053)-.003 .010 -. 172**(.033) (.024) (.054)- 191** . 058* . 054(.029) (.027) (. 058).045 . -.008 . 100
(. 051) (. 037) (. 073)-.012 .026 .061(.028) (. 020) (. 040)
E .005* .0007 .0001(.003) (. 007) (.005)
. 078* . 028 . 179**(, 036) (. 026) (. 054)
2
J4
J5
ct
C2
C3
C4
C5
R1
R2
H1
PSC1
PS1
p1
55. 30**(2. 33)- . 36**( . 09)2. 36**
( . 67)3. 55**
(1.01)3. 83**
( . 88)3. 29**
( .88)2. 85
(1. 79)3. 50 **
( . 96)8. 13**
(1. 13)8. 00**
(1. 88)9. 71**
(1. 28)9. 93**
( . 91)-2. 74**( .77)-1.34( . 81)-7. 35**( .73)1.40
(1. 24)-2. 00*( .69)
.02( . 06)2. 30*
( .90)
54.02'*(2. 84)
. 16
( . 12)1. 43
( . 83)1. 94
(1. 20)1.08
(1. 03). 20
( . 97)1. 36
(2. 03)2. 72*
(1. 09)1. 30
(1. 20)7. 15**
(2. 03)7. 38**
(1. 34)7. 24**
( . 96)-3. 41**( 92)-2. 95 **( . 96)
. 59(1. 10)
. 23(1. 47)-.10( . 80)-.10( . 10)2. 40*
(1. 03)
60. 69**(4. 02)
. 45**( . 16)-. 69
(1. 19)3. 04
(1. 67)2. 04
(1. 44)-. 60
(1. 33)5. 70
(3. 09). 81
(1. 50)3. 66*
(1. 70). 87
(3. 02)3. 05
(1.91)1.45
(1. 35)-. 99
(1.47)-. 24
(1. 51)1. 51
(1. 62)2. 27
(2. 03)1.64
(1. 12). 04
( . 13)5. 07**
(1.49)
Table 5-2, continued
W
W
2
3
179
.002 -.012 -.095* . 66 -1. 43 -2. 76**(.028) (.021) (. 043) ( .68) ( . 84) (1. 20).029 -.011 -.048 .64 -1.51 -2.03
(. 030) (. 023) (. 047) ( . 74) ( . 91) (1. 29).155 ** .042 -.086 2.95 ** - 2.62' -3.86*
(. 040) (. 029) (. 057) ( . 97) (1. 15) (1. 59)
* = . 05 level of significance4* = . 01 level of significance
180
test is .334 points greater than that of a person falling into
category Co, other things equal: the probability of a person in
category H1 passing the test is . 191 points less than that of a
person the same in all other ways but falling in category H0.
The numbers in parentheses under each estimate give the
standard error of the estimate. The * or ** indicates that the
estimate is statistically significant at the . 05 and .01 levels
respectively.
Education. - We see that attending junior college has a
positive effect over not attending junior college (or college) on
the probability of passing the MA, ranging in magnitude from
.083 (J1) to . 134 (J4). All the coefficients except J5 are
statistically significant at the 1 percent level of confidence and
increase in magnitude as years of education increase. Suppose
that the probability of success for a person in the reference
group in all categories (i. e. , Jo, Co, R0, H0, PSC0, P0. W0)
was . 255; if he remained in the reference group in all other
categories but moved to 34, his probability of success would
jump by . 134 to .389.
All the coefficients of the college variable are positive,
statistically significant at the 1 percent level, and higher than
any of the junior college coefficients, ranging in magnitude from
181
. 180 to .347. If the probability of success for a person in all
the reference groups was . 255, if he remained in the reference
category everywhere else but moved from Co (no college) to C5
(B. A. or B.S.), his probability of passing would increase to .602.
Thus, while college is not required for Oakland Police
Department applicants, it significantly increases the probability
of passing the MA by large magnitudes.
MA in column 4 of Table 5-2, translates these effects
into absolute increases in score. All educat ion coefficients but
J5 are statistically significant at the 1 percent level; junior
college attendance would increase the raw score between 2.36
and 3.83 points more than it would be without attending junior
college; senior college attendance would increase the raw score
from 3.50 (CI) to 9.93 (C5) points over what it would be without"
attending college.
Residence. - Living in the Oakland area or elsewhere in
California rather than out of the state has a weak (i. e. , not
statistically significant) negative effect on the probability of
passing the MA; living in the Oakland area has a significant
effect, however, on the score, decreasing it by 2.74 points.
These results may represent an interaction between positive
and negative effects of local residence discussed above. The
182
differences between the probability and score effects may be
partially explained by the fact that sometimes favorable adminis-
trative decisions may be made in favor of an occasional Oakland
resident, particularly of a minority group, allowed to proceed
through the testing process at a lower score than normally
considered passing.
Attending high school in the South. - Attending high school
in the South is significantly different at the 1 percent level from
attending high school in any other U. S. region. It lowers the
probability of success by .191 and the score by 7. 35 points. It
is probably measuring the combined effects of a high relative
number of blacks in the category and generally inferior southern
high schools.
Specific police training or experience. - Having attended
non-college police school has a small, positive, not statistically
significant effect on the probability of success and t'e score.
Having studied police science in college has a negative but
not statistically significant effect on the probability of success
and a negative statistically significant effect on the score,
decreasing it by two points. The negative sign is explained by
the fact that almost 90 percent of the police science students
come from junior college rather than college.
183
Job experience as a policeman has, as predicted, a positive
statistically significant effect on both the probability of passing
(. 078) and the score (2. 30).
Previous wage. - Only the highest wage category, $700 or
more per month, is significantly different from $0-$300 per
month; it increases probability of quccess by . 155 and score by
2.95 points. A salary of $700 a month or above in the period
1965 to 1971 indicates a professional or managerial job or success
as a salesman, insurance agent, independent proprietor, or
skilled craftsman. It is not surprising, therefore, that it is
indicative of.greater ability in passing the MA than a lower
salary. It is interesting that salary becomes important as an
explanatory variable only at such a relatively high threshold
rather than in continuous gradations. This may partially reflect
an imperfectly defined reference group which may include within
it part-time workers who are college students, for example. It
may also be an interesting reflection of the imperfect nature of
the American wage and salary structure.
Age. - Age, a continuous variable, has the predicted
negative effect; an increase of one year of age would decrease
the probability of passing by 015 and the score by .36. The
coefficients are statistically significant and small in magnitude.
The mean age of the applicants is 23 years, 10 months.
184
Number of jobs enumerated. - Number of jobs enumerated,
also a continuous variable, has a statistically significant, very
small (. 005) positive effect on the probability of passing and a
statistically insignificant effect on the score. The mean number
of jobs enumerated is 3.77.
Civil Service Test
One thousand 375 applicants successful on the MA took the
Civil Service test; 1, 266 or 92.1% of them passed it. Therefore,
there is not much variance in performance to explain and we
would expect few statistically significant coefficients. Only
three of the education variables now stand out from their reference
groups: having more than two years of junior college increases
the probability of passing by .064 over no junior college; having
three to four years of college increases the probability of passing
by .102 over no college; and having a B.A. or B.S. increases the
probability of passing by .081 over no college. In the score
equations, none of the junior college variables are significantly
different from no junior college. All the college variables except
C have significant effects on the score, with G3' C4 and C5 all
increasing the score more than 7 points above Co. C4 and C5,
so important to MA performance, are still differentially important
here.
185
The residence variables have no significant effect on the
probability, but decrease the score by 3. 41 (R1) and 2.95 (R2)
respectively.
Attending high school in the South is no longer significant,
having already weeded out those it would. Age, number of jobs
enumerated, PSC1 and PSI are also not significant. P and WI
W1
are not significant for probability but add more than two signifi-
cant points apiece to score.
Oral Exam
Between the Civil Service test and the Oral is the Physical
Agility, which will be discussed briefly later. Of 1, 093 people
taking the oral exam, 607 or 55.5 percent passed. Their
categorical distribution is summarized in Table 5.3.
The oral exam is conducted by a board of three: two
police officers, one Civil Service representative. Its philosophy
has been changing from an emphasis on selection for command
presence, verbal ability, logical thought processes, moral
feelings and social awareness to include knowledge about Oakland,
its problems and its minority population. The oral has not
changed in its attempt to screen as well for responsibility,
stability, reasonableness and some understanding of the function
of the job. Since the exam is set up to test for qualities more
186
Table 5-3 Percent of applicants in each category successfulon the oral exam
Category
TotalNumber
Examined
TotalNumber
Successful
Percent ofExamined
Successful
Total 1, 093 607 55.5J1 202 102 50.5J2 79 48 60.8J
3 130 76 58.5J4 164 85 51.8J5 21 15 71.4C1 96 56 58.3C2 72 43 59.7C3 21 12 57.1C4 58 43 74.1C5 136 78 57.4R1 597 320 53.6R2 367 189 51.5HI 80 51 63.8PSC1 49 35 71.4PSI 248 153 61.7
114 87 70.2wP1I 390 195 50.0W2 334 195 58.4W3 159 88 55.3
amorphous than economic variables can define, we would expect
our variables to have less explanatory power for oral exam per-
formance than for the original written screening.
There is no particular reason to expect, for example, that
education would be correlated with the above qualities, particularly
since that would imply an effect over and above the one already
exercised in allowing the applicant to reach the oral stage. None
187
of the junior college categories are significant; £4 has a signifi-
cant positive effect of .143 on the probability of passing; C2 has
a significant positive effect of 3. 66 on the score.
Oakland area residence has a significant negative effect of
.124, other California residence a significant negative effect of
.172, on the probability of passing. Neither has a significant
effect on the score. The considerably increased investment
required of an out-of-state resident to get to the oral stage is in
itself an indication of some of the qualities sought by the oral
and helps to explain the results. This may be offset in the 1970's
by a shifting emphasis to knowledge of Oakland and its problems.
Also, as would be expected, previous employment exper-tience as a policeman has a significant positive influence on both
success and score, increasing the probability of success by .179
and the score by 5.07 poims, the largest significant coefficients
found on any explanatory oral variables. Experienced policemen
may have already been screened for similar qualities to obtain
their previous jobs; they would have had oral exam experience;
they certainly could be expected to be more likely to exhibit
command presence, verbal ability, social awareness, respons-
ibility, stability, reasonableness and a superior understanding
of the job.
The salary variables behave erratically, with negative
coefficients in relation to the reference group significant for
W1 on both the probability and the score and for W3 for the
score. The W1 effect is reasonable given the fact that oral
boards are suspicious of people who would experience a large
jumb in salary by becoming policemen, and the reference group,
particularly at this point, may contain some college students or
others earning part-time salaries. The W3 score coefficient is
strange, particularly since its probability coefficient is positive
but not significant.
There is an interesting change in the sign of the age
variable, negative for the standardized written test, now becom-
ing positive and significant for the oral score. This offers con-
firmation to the desire of the oral board to screen for some
evidence of maturity.
Physical Agility
For curiosity, and because the data were available, we
ran the probability of passing the physical agility test against
the same explanatory variables plus height and weight. Out of
1, 270 taking the exam, 1, 093, or 86 percent, passed it. Age
had a negative effect of . 011 on the probability of passing,
statistically significant at 1 percent. Height had a negative
189
effect of .007 on the probability of passing, statistically signifi-
cant at 5 percent. None of the education variables were signifi-
cantly different from high school graduation except possession
of the A.A. degree, which had a negative effect of .133. Most
interesting is that residence in the Oakland area had a positive
effect of . 253 (significant at 1 percent) and residence elsewhere
in California a positive effect of . 227 (significant at 1 percent)
over out-of-state residence. Oakland and California residents
may be more familiar with the requirements of the test, which
are publicized in the announcement; they may have access to
gymnasium facilities in which to practice; and it is significantly
easier for them to return and re-lake the test should they fail
it the first time.
The only other significant variable was W3 , positive, . 095,
perhaps indicative of higher drive and motivation to succeed in
higher income earners.
1971 Equations, Includin: a Minorit Status Variable
The 213 applicants in the first few months of 1971 for whom
minority-majority status data were available were distributed
among the categories as shown in Table 5-4.
It is immediately apparent that minority group members
have a much lower percent passing the Mental Adaptability,
190
Table 5-4. Percent of 1971 applicant3 in each non-referencecategory successful on the Mental Adaptability
Category
TotalNumber
Applicants
TotalNumber
Successful
Percent ofApplicantsSuccessful
Total 213 162 76. 1
137 26 70.3
J?, 13 10 76.9
J3 32 25 78. 1T4 41 37 90. 2J5 11 7 63. 6C1 16 15 93. 8C2 16 15 93.8C3 1 1 100. 0
C4 6 5 83.3C5 25 23 92.0P1 131 92 70.2R2 73 63 86.3H1 24 8 33.3PSC 1 9 8 88.9PS1 48 38 79. 2P1 12 11 91.7
WI 62 48 77.4
W2 63 47 74. 6
W3 51 39 76. 5
MI 44 21 47. 7
47. 7, than average, 76.1, and than any non-reference category
except attended high school in the South, which picks up some of
the same effects.
It is of interest also to compare the percent of total
applicants in each category for whites and minority members
(Table 5- 5).
191
Table 5-5. Percent of white and minority members in eachnon-reference category
CategoryWhite Minority
Number Percent Number Percent
Total 169 100.0 44 100. 0,T1 30 17.8 7 15. 9J2 10 5. 9 3 6. 8J3 26 15.4 6 13. 6
J4 33 19.5 8 18. 2J5 8 4. 7 3 6. 8C1 14 8.3 2 4.5C2 15 8.9 1 2.3C3 1 0.6 0 0.0C4 3 1. 8 3 6. 8C5 20 11.8 5 11. 4R1 94 55.6 37 84. 1R2 67 39.6 6 13. 6H1 10 5. 9 14 31. 8PSC1 8 4. 7 1 0. 2PSI 37 21.9 11 25. 0P1 9 5.3 3 6.8
W1 52 30.8 10 22. 7W2 49 29.0 14 31. 8W3 41 24.3 10 22. 7
White applicants are 80 percent, minority applicants are
20 percent of the total number of applicants. Of the white
applicants, 5. 3 percent, and 13. 7 percent of the minority applicants
had no higher education beyond high school graduation, a difference
less than one might expect from their differential performance.
Otherwise, particularly at the highest levels, C4 and C5, the
educational distributions are surprisingly similar. A higher
192
percentage of minority applicants (84.1) than majority applicants
(55. 6) come from the local area. It is most interesting, from the
point of view of our 1965 to 1971 analysis, that 31. 8 percent of
the minority applicants and only 5.9 percent of the majority
applicants attended high school in the South. This creates a pre-
sumption in favor of our assumption that H1 was behaving partly
as a proxy for minority status.
We ran the same multiple regression equations for the 213
1971 applicants as we did for the 2, 366 1965 to 1971 applicants,
adding M1 = member of a minority group. With such a small
number of observations and so many cells, it is difficult to get
statistically significant coefficients, but for those which are
significant, the interpretation can be very strong. The results
are reported in Table
The most significant observations are in regard to the
behavior of the minority variable, M1. Being a member of a
minority group, other things equal, lowers the probability of
success on the Mental Adaptability by . 229 (statistically
significant at the 1 percent level) and lowers the score by 11.70
points (statistically significant at the 1 percent level). This is a
decisive and disturbing conclusion. Hi, which we believe is
co-linear with M1, also lowers the probability of success on the
193
Table C-6. Regression coefficient January to June, 1971,applicants
(1)MA
I
(2)CST
(3)OR
(4)MA2
(5)CST2
(6)OR 2
K 1. 056** 1. 14** .772** 6 2. 78** 36. 69** 68. 31**( . 283) ( . 235) (. 384) ( 8. 36) ( 4. 74) (11. 15)
A -. 021* -. 008 .049 57 -.07 - .05( . 010) ( . 009) (. 014) ( . 30) ( . 17) ( . 42)
J1 -. 054 . 081 . 018 1. 55 2. 44 . 34( . 079) ( . 06 4) (. 107) ( 2. 33) ( 1. 30) (3. 10)
J2 -.018 . 010 . 312* 1.31 2. 01 9. 23*( . 117) ( . 093) (. 156) ( 3. 46) ( 1. 88) ( 4. 54)
J3 .001 . 044 . 134 2. 53 1. 59 5. 39( . 083) ( . 06 5) (. 111) ( 2. 46) ( 1. 32) (3. 22)
J4 .089 . 105 . 187 5. 07* 2.64* 6.59*( 079) ( 059) (. 097) ( 2. 34) (1. 20) ( 2. 81)
J5 -. 192 -. 050 . 56 2** -2. 20 1. 45 10. 53( . 130) ( . 07 5) (. 205) ( 3. 84) ( 2. 32) ( 5. 95)
C1 . 130 -. 015 -. 125 4. 39 . 55 -1. 16( . 105) ( . 075) (. 124) (3. 10) (1. 51) (3. 61)
C2 . 196 . 016 -. 169 9. 96** - . 82 -2. 68( . 108) ( . 079) (.128) ( 3. 18) ( 1. 60) ( 3. 73)
C3 .322 . 073 . 091 I 2. 58 1.30 8. 66( . 413) ( . 288) (. 455) (12. 18) ( 5. 80) (13. 22)
C 4 . 139 . 107 -. 130 12.92 ** . 77 -4. 08( . 166) ( . 128) (. 252) ( 4. 91) ( 2. 59) ( 7. 32)
C5 . 202* . 083 . 126 7. 54** 1.08 2.47( . 090) ( . 066) (.1 08) ( 2. 6 5) ( 1. 34) ( 3. 14)
R1 .081 -.071 -. 666 ** -3. 70 -3. 82 -10. 36( 149) (.119) (. 19 1) ( 4. 39) ( 2. 40) ( 5. 53)
R2 . 129 -. 057 - .744 ** -1.65 -4.08 - 13.98*( . 155) ( . 123) (. 196) ( 4. 57) ( 2. 47) ( 5. 69)
H1 -. 314** -. 035 -. 020 -8. 16** -1. 13 - . 25( . 092) ( . 10 2) (. 174) ( 2. 71) ( 2. 05) ( 5. 06)
PSC1 . 094 -. 041 .302 6. 12 -.33 4.68( . 139) ( . 101) (. 173) ( 4. 08) ( 2. 03) ( 5. 04)
PS 034 -. 030 .060 -1.70 -.66 1.95( . 072) ( . 057) (. 094) ( 2. 13) ( 1. 15) ( 2. 73)
E 026 -. 00 4 -.0002 1.03 . 6 2* . 57
( . 020) ( . 015) (.026) ( .58) ( .31) ( .74)Pl 208 . 096 .354* 3.02 2. 13 7.08
( . 127) ( . 09 2) (. 151) ( 3. 74) ( 1. 86) ( 4. 38)
194
Table 5-6, continued
W1 . 046 . 024 -. 116 . 04 -. 50 2. 81(. 083) (. 065) (. 106) (2. 46) (1. 30) (3. 08)
W2 . 051 . 005 . 020 4. 35 -. 04 4.64(. 084) (. 065) (. 109) (2. 47) (1. 32) (3. 15)
W3 . 075 . 046 -. 053 1. 45 . 09 4. 06(. 099) (. 079) (. 128) (2. 93) (1. 59) (3. 72)
M -. 229** -. 0291. 515** -11. 70** -3. 78** 10. 76**
(. 073) (. 067) (. 115) (2. 14) (1. 35) (3. 35)
* = . 05 level of significance** = . 01 level of significance
195
MA by . 314 points (significant statistically at 1 percent) and the
score by 8.16 points (significant statistically at 1 percent). The
probability equation for the Civil Service test has no statistically
significant results; on the score equation minority status lowers
the probability of success by 3. 78 points (statistically significant
at 1 percent).
On the oral, the effect of minority status is reversed,
however, becoming positive and raising the probability of passing
by . 515 and the score by 10. 76 points (both statistically significant
at 1 percent). This is an explicit reflection of a policy attempt to
hire more minority members, but the attempt begins only after
successful performance on the written. On the oral, the inclusion
of knowledge of Oakland, its minority community, and its problems,
appears to assist in the performance of the minority candidates.
These results, however, underscore the critical importance of
careful evaluation and interpretation of the use of standardized
written exams and their results for selective purposes, particu-
larly where it is essential and desired to increase minority
representation.
The behavior of all other variables simply bears out, in
much weaker fashion because of the small number of observa-
tions, the conclusions already discussed in the more general
analysis, and therefore won't be belabored here.
196
CHAPTER
THE CHALLENGES OF THE 1970's
This chapter will discuss the extremely significant policy
changes, particularly in the minority composition of the force,
between 1971 and August, 1974. It will raise, as well, some
further economic policy questions for the Oakland Police Depart-
ment to consider, particularly about options that may be open to
revise economic incentives to attract and retain the type of labor
force it desires.
MINORITY REPRESENTATION
According to the 1970 Census, the City of Oakland is over
one-half minority. The population is 34. 5 percent black, 9.8
percent Spanish surnamed, and 6. 5 percent other minorities. 1
Yet, in 1966, just 3 percent of the Oakland police force was
minority; by the end of 1970, the percent minority had increased
to 6. 5. 2 The relationship between the under-representation of
the minority communities on the force and the apparently
discriminatory selection instruments discussed in Chapter 5
was clear to the minority communities.
The Legal Environment
On October L7, 1969, a suit was filed in the United States
197
District Court, Northern District of California, the case of
Clarence Penn vs. William Stumpf, Civil Action No. C-69-239
OJC, "seeking equitable relief to remedy an alleged pattern and
practice of discrimination in hiring for the Oakland Police
Department" violating civil rights. 3 The complaint, which was
more than 20 pages long, generally alleged that the processes of
selecting applicants for appointment for positions with the Oakland
Police Department unlawfully discriminate against the plaintiff
and other minority, i. e., black and Spanish surnamed persons,
in the following ways:
1. By the use of a written examination which:
a. is not practical in nature,
b. does not fairly test the fitness of the person
examined to discharge the duties of a police
officer,
c. measures skills substantially unrelated to
that position,
d. requires linguistic, verbal and other facilities
unrelated to the position,
e. discriminates against Negro and Spanish
surnamed minority group males solely on
account of their race, education and racial
198
and cultural backgrounds,
f. has never been subjected to an impartial
professional validity study,
g. lacks predictability of success,
h. has kept worthwhile qualified minority males
out of the Department.
2. By using a background investigation as a screening
method which investigation:
a. uses as screeners persons who are themselves
unfamiliar with minority life styles and
cultures,
b. uses Caucasian officers to screen who have
not been educated or trained for that purpose,
c. uses Caucasian officers who themselves are
unscreened as to the existence of racial
prejudice on their part. From the screening
process information is obtained about appli-
cants and used to bar them from the position
despit,.1 the fact that such information is com-
pletely unrelated to the question of whether the
applicant is capable of performing the duties
of an officer hi the Oakland Police Department
199
and therefore serves to keep out minority
members.
3. By using recruitment policies which:
a. utilize white personnel who have little or no
contact with potential minority group police
officer applicants,
b. recruit at schools and other institutions where
the vast majority of potential applicants are
Caucasian, and thereby
c. prevent minority group males who might be
potential applicants from learning of openings.
4. By the use of a psychiatric and oral examinations
which:
a. are conducted by persons unfamiliar with the
culture and life styles of black and Spanish
surnamed societies,
b. are conducted by persons who use medical
and mental frames of reference appropriate
to white middle class culture as a result of
which the plaintiff and other minority members
are discriminated against.
5. With respect to the oral interview, it is alleged
200
that this interview is conducted by Caucasians who
are unscreened for their familiarity with minority
groups and who either harbor feelings of race
prejudice or are ignorant of the history, culture
and background of minority group persons so that
they are unable to make valid judgments as to
qualifications of individuals from those groups. 4
On February Z, 1970, the court denied a motion to dismiss
filed by the city and held that the plaintiffs' complaint stated a
cause of action. 5 The City of Oakland filed an answer on Feb-
ruary 11, 1970, which denied all material allegations.
On March 8, 1971, it became apparent that changes in
selection devices or results would have to be made. On that
date, the United States Supreme Court rendered its unanimous
decision in the landmark case of Willie S. Griggs v. Duke Power
Company, Z8 L ed 2d 158, brought under Title VII of the Civil
Rights Act of 1964 (42 U.S. C. 2000e, et. mg.). The Supreme
Court ruled that if the use of certain tests and educational re-
quirements for screening job applicants results in a higher re-
jection rate for blacks than for whites, the employer must
demonstrate that passing such tests is significantly related to
successful performance on the job. "If an employment practice
201
which operates to exclude Negroes cannot be shown to be related
to job performance, the practice is prohibited. "6
The Equal Employment Opportunity Commission's Guide-
lines on Employee Selection Procedures 7 underscores the re- .
quirements more clearly:
For the purpose of the guidelines in this part, the term
"test" is defined as any paper-and-pencil or performance
measure used as a basis for any employment decision.
The guidelines in this part apply, for example, to
ability tests which are designed to measure eligibility
for hire, transfer, promotion, membership, trait.ing,
referral or retention. This definition includes, but is
not restricted to, measures of general intelligence,
mental ability and learning ability; specific intellectual
abilities; mechanical, clerical and other aptitudes:
dexterity and coordination: knowledge and proficiency;
occupational and other interests; and attitudes pel .on-
ality or temperament. The term "test" includes all
formal, scored, quantified or standardized techniques
of assessing job suitability including, in addition to
the above, specific qualifying or disqualifying personal
history or background requirements, specific educational
202
or work history requirements, scored interviews,
biographical information, blanks interviewers' rating
scales, scored application forms, etc. 8
Discrimination is defined as "the use of any test which
adversely affects hiring, promotion, transfer, or any other
employment or membership opportunity of classes protected by
title VII" 9 unless "the test has been validated and evidences a
high degree of utility. "10 The protected classes are defined by
race, color, religion, sex, or national origin. "11 "Evidence
of a test's validity should consist of empirical data demonstrating
that the test is predictive of or significantly correlated with
important elements of work behavior which comprise or are
relevant to the job or jobs for which candidates are being
evaluated. "12
Moreover, empirical evidence to support a test's validity
must be based on generally accepted procedures, approved by
the American Psychological Association, 13 for determining
criterion-related validity. Where criterion-related validity is
not feasible, evidence of content or construct validity may be
acceptable, if "accompanied by sufficient information frOm job
analyses to demonstrate the relevance of the content (in the case
of job knowledge or proficiency tests) or the construct (in the
203
cast: of trait measures), Evidence of content validity alone may
be acceptable for well-developed tests that consist of suitable
samples of the essential knowledge, skills or behaviors composing
the job in question, " not including "those which can be acquired
in a brief orientation to the job. "14
It became amply clear from the developing case law and
from the Equal Employment Opportunity Act of 1972 (P. L. 92-261,
86 Stat. 103) which extended the Civil Rights Act of 1964 to public
employers that the City of Oakland could be forced by the courts
to answer two questions: 1) does any selection device have a
discriminatory adverse effect on the hiring of minorities? , and
2) if so, can that selection device be shown to be job related?
If a selection device does have adverse effect, even if it is job
related, the employer must also show that he knows of no alter-
native technique that would have less of an adverse effect on
protected groups and is equally valid. 16
Minority Composition of the Oakland Police Department
Adverse effect, somewhere in the Oakland recruitment and
selection process, seemed clear, particularly until 1971, from
the statistics on the percent of new appointees to the Police
Department that were minority group members. Table 6-1 pre-
sents the statistics on the percent of new appointees that were
204
minority group members from 1961 to 1973. Note that the per-
centages doubled in 1971 over 1970, and doubled again in 1973 over
1972 to a dramatic 61 percent of new appointees from minority
groups. Policy changes had to be at work:
Table 6-1. Percent of new appointees from minority groups,Oakland Police Department, 1961 to 1973.
YearNumber of
New AppointeesNumber from
Minority Groups
Percent of NewAppointees fromMinority Groups
1961 24 1 4.11962 44 5 11.31963 49 3 6.11964 43 0 0.01965 61 3 4.91966 81 6 7.41967 49 7 14.21968 91 16 17.61969 69 14 20.31970 73 11 15.11971 85 30 35.31972 94 28 29.71973 72 44 61.0
Source: Oakland Civil Service Office
The changes in the total number of Police Department
positions filled, over all ranks, by members of the various
minority groups, is shown from December, 1966, to December,
1973, in Table 6-2. At the end of 1966, just 3 percent of the total
Tab
le 6
-Z
.T
otal
num
ber
and
perc
ent o
f O
akla
nd P
olic
e D
epar
tmen
t pos
ition
s fi
lled
by e
ach
min
ority
gro
up,
Dec
embe
r, 1
966,
to D
ecem
ber,
197
3
Dat
e
Tot
alFi
lled
Posi
-tio
nsB
lack
sSp
anis
hSu
rnam
eA
sian
Am
eric
anN
ativ
eA
mer
ican
Tot
alM
inor
ityN
umbe
r Pe
rcen
tN
umbe
rPe
rcen
tN
umbe
r Pe
rcen
tN
umbe
r Pe
rcen
tN
umbe
r Pe
rcen
t
Dec
embe
r,19
6666
115
2.27
40.
601
0.15
00.
0020
3.02
Dec
embe
r,19
6764
818
2.77
40.
611
0.15
00.
0023
3.55
Dec
embe
r,19
6865
820
3.03
40.
601
0.15
10.
1526
3.95
Dec
embe
r,19
6968
228
4.10
71.
020
0.00
10.
1536
5.27
Dec
embe
r,19
7070
534
4.82
101.
411
0.15
10.
1546
6. 5
3D
ecem
ber,
1971
714
385.
3211
1.54
40.
563
0.43
567.
85D
ecem
ber,
1972
700
557.
8615
2.14
131.
852
0.30
8512
.15
Dec
embe
r,19
7369
571
10.2
122
3.17
162.
303
0.43
112
16.1
1
Sour
ce: O
akla
nd C
ivil
Serv
ice
Off
ice
206
force was minority: by the end of 1973, 16 percent of the force
was minority.
Of the men on the force in October, 1973, 19. 2 percent of
the Caucasians were hired after 1971; 56.3 percent of the blacks,
57.1 percent of the Spanish surnamed, and 70.6 percent of the
other minorities represented (Asian-Americans, Filipinos,
Hawaiians and Native Americans) were hired after 1971. For a
detailed breakdown of the force, by ethnic groups and year of
hire, see Table 6-3.
Partly because the minority group members have come on
the force in numbers only since 1971, there are only four minority
officers, all blacks, above the rank of patrolman. Two are ser-
geants, one is a lieutenant, one a deputy chief.
Administrative Attitudes
What was going on in 1971, 1972 and 1973 to produce such
dramatic changes? It should be clear that Oakland's Police
Chief, Charles Gain, and Personnel Director, James M. Newman,
both superior professionals, recognized the urgency of drawing
in more minorities and actively supported research and analysis
of existing procedures and possible changes therein. The legal
developments, in fact, seemed almost to provide them with the
necessary political leverage, against more obstructionist groups
207
Table 6-3. Oakland Police Department members, by date ofappointment and ethnicity, October 30, 1973
Number NumberYear of Number Number Spanish OtherAppointment Caucasian Black Surnamed Minority Total
19401941
1 1
0
1942 2 2
1943 2 2
1944 1 1
1945 6 1 7
1946 5 5
1947 25 1 26
1948 4 41949 28 2 301950 4 4
1951 27 1 28
1952 3 1 1 5
1953 lt 1 17
1954 ?.0 10
19C; 9 9
1956 13 13
1957 19 19
1958 9 9
19 59 ) 9
1960 1 1
1961 3 3
1962 9 2 11
1963 8 1. 9
1964 17 17
1965 24 1 25
1966 37 Z 1 401967 28 1 1 301968 48 7 2 57
1969 58 6 2 1 671970 50 7 3 601971 44 10 4 2 60
1972 44 10 2 7 631973 25 20 6 3 54
Total 589 71 21 17 698
Percent of force 84. 4 10. 2 3, 0 2. 4
Source: Oakland Civil Service Office
208
on the police force and in the community, to institute policies
which would work in the right direction. They knew, and could
argue persuasively, that if they could not voluntarily design ways
to hire more minorities and improve the statistics, they would
soon be under much less palatable court orders.
Both men were extremely sensitive, as well, to the need
to preserve high quality standards. Admitting inferior candidates
would pose unacceptable risks both to the quality of the police
force and to the effort to bring in more minorities. They were
sophisticated, however, about the difficulties of choosing
appropriate selection instruments to measure quality. Through-
out the 1960's they had encouraged interested researchers to use
departmental data to try to relate selection devices to job per-
formance. The results were inconclusive. If adverse effect
were demonstrated in court, it would be very difficult to prove
that selection devices, particularly the written tests, were job
related. It might be easier to remove the adverse effect.
Research Efforts
There are several important barriers to demonstrating
conclusively through empirical research that a test discriminates
well between good and bad job performers. Foremost among
these difficulties is that selection test failures are not given an
209
opportunity to show whether or not they can perform on the job.
Therefore, there can be no observations about the job performance
of failures. Supervisors' evaluations show very little variance;
paired comparisons are time-consuming and often unreliable:
turnover data tells little about performance; and, particularly
in a force the size of Oakland's, there may be too few performers
at either extreme to yield statistically significant conclusions.
William K. Hunter and Thomas J. Culligan, for example,
in an attempt to identify which factors could be used to predict
police performance, asked police supervisory personnel to
compare and evaluate prior subordinates who had subsequently
left the lepartment, on a forced-choice paired comparison basis. 17
They used people who had left the department in an attempt to
reduce many potential biases from work unit associations and
active friendships. From an original group of 146 policemen
who had left the department, 120 were chosen who had sufficient
time on the force and sufficient experience with supervisors.
Then, the sample to be studied was confined to the top 28 and
bottom 28 individuals as ranked. 18
The result of most interest to us here is that as a group,
they don't do significantly differently on the written test. The
average score of the top 28 on the Otis Self-Administering Test
210
of Mental Ability was 54. 5. The average score of the bottom 28
was 53. 2. An F-test shows that the difference is not statistically
significant. 19 On the other hand, the high scorers were definitely
more capable of handling the Civil Service oral examination.
The average oral score of the top 28 was 84. 75, of the bottom
80. 0. An F-test shows a statistically significant difference at
a better than 1 percent level of confidence. 20
In William K. Muir's in-depth, but small, study of 27
Oakland policemen, we find that the intelligence tests, particu-
larly the Army General Classification Test, do appear to be use-
fullyfully related to the best professional performance. Muir
argues that this is because they test verbal experience, and "the
professionals' enhanced verbal experience eases the task of
coping with the social cross pressures of the job, of resolving
the dilemma of power, of developing a philosophy, and of being
a leader within the department. '122
Muir also finds that the Civil Service general knowledge
test is useful. Muir's "impression is that the narrower the
experience and the less diverse the social grown with which the
candidate has warm familiarity, the worse he ctcores',23 and that
"a poor performer on the CSW ought to be marked as a candidate
with real inadequacies in his familiarity with the world. "24 He
211
believes that a preannounced current affairs test should be
included in the testing battery, particularly as an opportunity for
a candidate to excel "if he has the motivation and minimal learning
skill to do so. "25
High value is also placed by Muir on the oral interview,
which "assesses motivation and maturity in a way no written test
can and is critical in the evaluating process. "26 Muir would be
the first to stress, however, that his findings are suggestive and
not statistically significant. They seem to caution against a too
hasty abandonment of traditional selection methods under court
attack.
Personnel Director James M. Newman summarizes the
research attempts as indicating that the Mental Adaptability and
Civil Service tests had no real predictive validity corresponding
to success on the street. The strongest demonstration that could
be made was that test scores were correlated with performance
in the police academy and that performance in the police academy
was correlated with later performance evaluations. 27 But no
direct correlation could ever be demonstrated between the test
scores themselves and performance, 28
A "Content Valid" Written Exam
By November, 1970, the beleaguered Civil Service test of
212
general knowledge and current events was replaced by a written
test basically of reading comprehension. The test has 50 ques-
tions; 30 must be answered correctly to pass. The first section
consists of 25 multiple choice questions related to six pages of
crime reports, to be answered in 40 minutes. The second
section consists of 25 multiple choice questions related to eight
pieces of narrative material, to be answered in 40 minutes. By
the middle of 1971, this test replaced the Mental Adaptability tests
as well. Therefore, neither the Mental Adaptability tests nor the
Civil Service test which we analyzed in Chapter 5 is currently
being used. The written reading comprehension test replaces
them.
The new written test has not been validated technically as
related to job performance. However, high scores have been
shown to correlate positively with high perforMance in the
recruit academy. It is argued, therefore, that the test is
constructed only to predict success in the recruit academy, in
which an applicant will deal with similar material. Therefore,
it is reasonable on its face and content valid.
The Affirmative Action Survey by the California FEPC
still is critical of the test as indicating "only one skill: the
ability to analyze, under pressure, difficult reading material. "30
213
According to the Fog Readability Index, the second part of the
examination is beyond high school reading level. 31
However, in the Consent Decree in the Penn v. Stumpf
case, signed December 20, 1973, it is agreed that: "The Civil
Service Board will continue its efforts to develop or obtain a
demonstrably job-related written test for the selection of police
officers. The present examination may continue to be used on a
pass-fail basis per:ding those efforts. "32
College seniors or college graduates with majors in
Administration of Justice, Law Enforcement, Police Science,
Criminology, Sociology, Psychology, Public Administration, or
related fields, are now able to request a waiver of the written
examination by submitting their transcripts for review.
The revision of the written exam has not removed the
adverse effect on minorities. As Table 6-4 shows, 84.7 percent
of the Caucasians taking the test between January, 1972, and
September, 1973, passed, compared to 41.6 percent of the
blacks, 66. 7 percent of the Spanish surnamed, and 73.0 percent
of the A3ian Americans. Of all waivers granted, 72. 4 percent
were to Caucasians, 20. 7 percent to blacks.
Yet, the exam now does have content validity and will be
retained provisionally until something better can be developed,
Tab
le 6
-4.
Exp
erie
nce
of e
thni
c gr
oups
on
patr
olm
an e
xam
inat
ion,
Jan
uary
,Se
ptem
ber,
197
319
72, t
o
Patr
olm
an A
pplic
ants
Cau
casi
anB
lack
Span
ish
Surn
ame
Asi
anA
mer
ican
Nat
ive
Am
eric
anT
otal
Tot
al e
nter
ing
test
ing
prog
ram
884
502
6145
81.
500
Perc
ent o
f to
tal
58. 9
33. 5
4. 1
3. 0
0. 5
100.
0
Num
ber
taki
ng w
ritte
n72
745
757
375
1, 2
83N
umbe
r pa
ssin
g w
ritte
n61
619
038
275
876
Perc
ent p
assi
ng w
ritte
n84
.741
.666
.773
.010
0.0
68.3
Num
ber
of e
xam
s w
aive
d15
745
48
321
7Pe
rcen
t of
ethn
ic g
roup
wai
ved
17. 8
9.0
6.6
17.8
37.5
14. 5
Perc
ent o
f al
l exa
ms
wai
ved
72. 4
20. 7
1. 8
3. 7
1. 4
100.
0
Tot
al r
emai
ning
aft
er w
ritte
n77
323
542
358
1, 0
93
Tot
al h
ired
7441
76
112
9
Perc
ent h
ired
9.6
17.4
16. 7
17. 1
12. 5
11. 8
Sour
ce: O
akla
nd C
ivil
Serv
ice
Off
ice
215
if possible. Recruits, we should note, must pass successfully
through a 22-week training academy after they are hired. The
academy emphasizes academic as well as practical knowledge;
for example, constitutional and criminal law, criminal evidence,
criminal investigation, community awareness, report writing,
and patrol procedures. Many of these subjects are taught by
college level textbooks. Basic verbal and reading comprehension
skills are essential for successful academic performance in the
academy, and, one might argue, might well be necessary on the
job.
Therefore, the Police Chief and Personnel Director are
very anxious to retain a written test as a quality screening
device. They would prefer to concentrate on attracting a suffi-
cient number of qualified minority candidates to enter the selec-
tion process so that acceptably large numbers will survive the
written exam. It would be interesting, as the data accumulate,
to test empirically whether the qualities needed to pass the
revised written exam are the same as or different from the
qualities needed to pass the Mental Adaptability and Civil Service
tests. Scope for future research is clearly indicated.
Other Changes in the Selection Process
Our study of the 1965 to 1971 period showed that minority
216
group status was no longer a disadvantage once a candidate
reached the oral exam. In fact, in the first half of 1971, minority
status raised the probability of passing by . 515 points and the
score by 10. 76 points (both statistically significant at 1 percent).
The positive effect of minority status on the oral continued
through 1971, 1972 and 1973. In the January, 1972, to September,
1973, period, 66. 2 percent of the Caucasian candidates failed the
oral; 44. 3 percent of the blacks, 47. 6 percent of the Spanish
surnamed, 54. 3 percent of the Asian Americans, and 50. 0 per-
cent of the Native Americans failed the oral. 33
Indeed, the Affirmative Action Survey by the California
FEPC states that "the adverse effect of the written is totally
dissipated by the oral. "34 Since the written examination is now
only qualifying, it is the oral score that determines a candidate's
position on the eligibility list. The oral board consists of three
members, an Anglo police captain who is the Police Department
Personnel Officer, a black sergeant, and a Civil Service repre-
sentative, also Anglo. The team remains consistent and constant
and has developed a high caliber interviewing technique, which
carefully probes a candidate's attitudes toward minority and low-
income groups. 35
According to the FEPC's assessment, the Physical Agility,
217
the. Medical Examination, the Psychiatric Examination, and the
Background Check do not yield any significant evidence of ad-
verse effect on minorities. The background investigation has
been improved recently by attempting to eliminate the use of
police officers on overtime. All investigations would then be
conducted by the two officers, representative of minorities,
assigned full-time to do background checks. All rejects must be
approved by the Police Chief and there is an appeal process.
As Table 6-4 shows, between January, 1972, and Septem-
ber, 1973, 9. 6 percent of the Caucasians who passed the written
were actually hired; 17. 4 percent of the blacks, 16. 7 percent of
the Spanish surnamed, 17.1 percent of the Asian Americans and
12. 5 percent of the Native Americans who passed the written were
actually hired. In actual numbers, this meant 74 Caucasians, 41
blacks, 7 Spanish surnamed, 6 Asian Americans, and 1 Native
American were hired in this period.
Active Recruitment of Minority Candidates
One way of hiring more minorities, even without making
changes in the selection procedures, is to attract more qualified
minority candidates to apply. The City of Oakland had always
devoted a great deal of attention to attracting qualified police
candidates. Indeed, it pioneered ambitious, nationwide, college
218
police recruiting efforts and visited a wide range of college
campuses in the 1960's. Nation.al recruiting campaigns were
ended in late 1969 as the result of a few unpleasant campus
demonstrations, mounting political pressure to intensify local
recruitment efforts, and indications that nationally recruited
candidates were more likely to leave the force. In Table 2-7
we showed, for example, that between August 1, 1962, and
August 1, 1971, 43.6 percent of the out-of-state residents re-
signed from the force, compared to 20. 2 percent of local resi-
dents and 19. 3 percent of other California residents.
On July 20, 1971, the pendulum had swung full cycle. The
residence requirement, which had been removed by Oakland
voters in 1956, was reinstituted. Open competitive examina-
tions for civil service appointments were restricted to city
residents, to the extent that resident applicants are available. "36
In practice, non-residents are allowed to take the police exams,
but they cannot be appointed until all residents on the list have
been appointe,:. If appointed, a declaration of intention to move
into Oakland within nine months after appointment is required.
Enforcement has been somewhat ambiguous, partly because of37some doubt about the constitutionality of residence requirements,
but a preference for residents is clear.
219
On October 26, 1971, the Oakland City Council adopted a
much heralded and publicized Affirmative Action resolution
which "sets a goal for eliminating any disparity between the mi-
nority composition of City employment and total City population,
and calls for parity to be achieved in all City departments at all
salary levels and in all job classifications. "38 The specific
implementation of the Affirmative Action program for the Oak-
land Police Department has been incorporated in the Penn v.
Stumpf Consent Decree and requires that:
A. For all minority persons whose employment with
the Oakland Police Department terminated for any
reason during each year, minority persons on the
eligibility list shall be hired to fill the vacancies
created within the same year period.
B. For all new hires other than those covered by sub-
section A, the percentage of minority persons
hired each year from the Oakland Police Depart-
ment's eligibility list shall at least equal the
current percentage of minority persons in the pop-
ulation of the City of Oakland. In carrying out this
procedure the Oakland Police Department shall
hire minorities from each ethnic group in general
220
proportion to the City's ethnic composition,
assuming an adequate representation of minorities
of each ethnic group on the eligibility list. 39
The Consent Decree specifies further that these are goals, not
quotas, and that no employees shall be dismissed to achieve the
goals. The hiring of unqualified persons who have not passed the
examinations and hiring requirements, or of persons within the
lowest 10 percent of the candidates on the eligible list, is not
required to achieve the goals.
If, at any time, the Oakland Police Department cannot hire
minority persons according to the goals because an insufficient
number of minorities have passed the initial hiring requirements,
then the department is required to proceed with continuous testing
procedures, make special efforts at recruiting minorities, and
notify the attorneys for the plaintiffs of the insufficiency and try
to explain the reasons. 41
Active recruiting is viewed as a primary prerequisite of
success in achieving minority hiring goals and maintaining
quality standards. Job opportunities are publicized through
local organizations, high schools, junior colleges and colleges
with substantial minority enrollments. Since December, 1972,
local college placement centers are visited and contacts are
221
made v,Itn faculty personnel and minority student counselors for
referrals of qualified prospects. After an applicant is referred,
the "one-on-one" approach is used--that is, an appointment is
made to meet him personally and provide individual attention
and information about the selection process. This approach is
viewed as more successful than meeting groups of potential
applicants, because the departmental representatives do not have
to be on the defensive against possibly hostile individuals or
negative peer group pressure.
In 1973, a media saturation program was launched, funded
by the Law Enforcement Assistance Administration through the
California Council of Criminal Justice. Four Bay Area depart-
ments, Oakland, San Francisco, Berkeley and Richmond, co-
operated in mass media advertising aimed at the minority popu-*
lace and designed by Anrich, Inc. of Oakland, a minority-owned
advertising agency. The campaign consisted of 30 and 60 second
spot commercials on television and radio, using real minority
police officers, billboards, posters on transit vehicles, etc.
Interested people were invited to call in to a telephone center,
manned by Oakland, for further information. Callers were con-
sidered eligible for recruitment only by the police agency in their
city of residence for 48 hours, after which he could be considered
222
by the other cities as well. Oakland tried to schedule an inter-
view with all interested callers to discuss in depth the advantages
and disadvantages of police employment, the selection process,
and the employment application and to arrange for interested
applicants to tour the police facility or to "ride along" with a
patrol officer. This continued the "one-on-one" approach to
rec ruiting.
Between April 24 and October 11, 1973, the telephone center
processed 2, 102 calls. Twenty percent of the callers were
apparently ineligible because of citizenship, age, height, weight,
or some other requirement: 76 percent of the eligible callers
lived in Berkeley, Richmond or San Francisco. Oakland received
490 calls from apparently eligible Oakland residents. Two
hundred and 34 individuals took the initial written test; 93, or
39.7 percent, passed the written test. Fifty-three people (57
percent of those who passed the written) passed the oral. Fifteen
blacks and four Mexican Americans were hired by the department
as a result of the Minority Recruitment program; it is projected
that a total of 36 minority persons will have been hired by Oak-
land by the project's end. As for the other participating depart-
ments, Berkeley received inquiries from 83 eligible candidates,
Richmond from 68 and San Francisco from 591, but what
happened to these candidates is not known. 42
223
The cost to Oakland, exclusive of $28, 368. 27 in Federal
funds, as of October 1, 1973, was $44, 524. 11, or $2, 120.19 per
minority person recruited. The Oakland Civil Service Office
estimates that in past years approximately $1, 000 per recruit
has been spent for each minority officer. If the projections of
36 minority persons hired are correct, the 1973 project will have
cost the Federal government $30, 618 and Oakland $68, 054, or
$1, 890.41 per recruited person for Oakland. 43
Most importantly, through all these public relations and
recruitment efforts, the public image of the department has
changed. It is now amply clear that the department wants minor-
ity candidates to apply. Moreover, minorities are seen on the
force in ever-increasing numbers, visible proof that they can
succeed in the selection process and be accepted as members of
the fcrce. The minorities hired in the past few years can refer
interested friends and relatives. In short, to quote Personnel
Director James M. Newman, "The more you have, the more
44you get. Sixty-one percent of the new appointees in 1973
were from minority groups. All passed through the selection
process and were deemed qualified. If this success can con-
tinue for the next few years, and there is every indication so far
in 1974 that it can, the Oakland Police Department's worries
224
about minority representation will be over. So will Oakland's
legal troubles on that front, since if the statistics do not bear
out adverse effect, individual selection devices do not have to be
validated as job related. Oakland's experience in 1973 demon-
strates clearly what can be accomplished with motivation and
effort.
WOMEN
Challenges have come, however, from another protected
class as well: women. Patrolman and policewoman had always
been two distinct categories of police officer. There were seven
policewoman positions in the department. The women were
assigned not to patrol duties but to the juvenile division, the
women's prison, rape cases, etc. Requirements were different
for policewomen and patrolmen. The women could be shorter
and lighter: 5'4". and 120 pounds, minimum height and weight.
They were also required to have either a college degree or four
years of relevant work experience. They were given the same
written examinations as the men but a different Physical Agility
designed especially for women. The rest of the hiring process
was the same. Whereas there was continuous recruiting and
testing of male applicants, policewoman positions came up
infrequently.
225
Two 1aws11:ts were brought, one by an applicant seeking a
patrolman's position, one by a policewoman seeing a promotion
to sergeant. Under legal pressure, on February 1, 1974, the
Oakland City Council merged the patrolman and policewoman
classifications into one new classification, police officer. 45
They set a standard height and weight requirement, 5'7", 135
pounds, minimum height and weight. The differences in educa-
tional requirements were dropped.
A class suit was brought against the height and weight
minima, arguing that they impermissibly discriminated against
women as a class. On March 15, 1974, the California Court of
Appeals struck down the city's height and weight requirements
in the case of Veragene Haidy vs. William F. Stumpf. 46The
Court argues that over 80 percent of all American women are
excluded by such a size limitation. The trial court's finding
that the size requirements are reasonable and necessary to the
normal operation of the police patrolman's duties is unsupported
by the type of evidence needed to justify discriminatory treat-
ment. The Court did not dispute that physical strength may be
related to performance, only that a person must be 5'7", 135
pounds to possess the necessary strength. In fact, Justice
Brown added this statement:
226
The police arm functions not only to protect the public,
but to instill in that public confidence that it is being
protect d--an increasingly difficult task in these days
of rising crime rates. Some non-discriminatory
minimum size requirement is entirely in order, and
to the extent that the opinion in this case may be deemed
to strike down all such limitation as a matter of law,47we cannot agree.
The Court invited the city to reexamine the evidence that
the requirement is discriminatory, based on local and recent
data, or to demonstrate conclusively that it is required. On
May 10, 1974, the California Supreme Court denied a hearing in
the case. The decision of the Court of Appeals, abolishing
minimum height and weight requirements, became final on.
April 15, 1974.
The city feels that without the height and weight require-
ments, the Physical Agility test is of doubtful use as a screening
device, because it w,,s designed assuming the minimum height
and weight of candidates. Testing was suspended and applica-
tions held in abeyance during the summer of 1974 to allow foi
the development of a new Physical Agility test applicable to the
new circumstances.
227
in the promotion case, Joan Hail vs. James S. White, a
policewoman alleges that she has been denied an opportunity to
become a sergeant, for which she was otherwise qualified,
solely because of sex discrirnination. 48 An officer must have
had three years of experience as a patrolman to be eligible to
take the sergeant's examination. On this basis, Joan HRH was
refused permission to take the exam. Finally, under pressure,
the city allowed her to take the written examination, which she
passed, and the oral, the results of which are sealed, until the
legal issues are settled. On August 2, 1973, the Federal
District Court upheld the city's requirements, arguing that they
are neither unreasonable nor non-job related, that Ms. Hail's
inability to meet the conditions for promotion is a product of hur
personal inability to meet the rcasonable requirements of patrol
49service.
The case is still in the C our t of Appeals, and the appeal
looks like it may well be successful. An Amicus Curiae brief
submitted by U. S. Assistant Attorney General J. Stanley
Pottinger argues that the District Court erred in employing a
standard of reasonableness rather than a standard of business
necessity for employment practices which have a discriminatory
50impact. Pottinger argues that Hail is concededly qualified to
228
be a police sergeant, except for her lack of experience as a
patrolman, her height and weight, and the fact that she never
took the Physical Agility test required of patrolmen. Requiring
that she have three years of experience as a patrolman needlessly
perpetuates the effect of past discrimination, since she was
denied that experience on the basis of her sex. Moreover, the
Oakland Police Department made no showing that three years as
a patrolman are necessary to do the work of police sergeant.
Not all sergeants have had patrol experience and most are not
assigned to patrol duties. Moreover, Ms. Hail had experience
similar to that of a patrolman, including patrol duties as a
policewoman. There are no height, weight or Physical Agility
requirements for promotion to sergeant, and the height and
weight requirements for patrolmen were struck down by Hardy
vs. Stumpf. The height, weight, and Physical Agility have not
been demonstrated to bo a business necessity. Therefore, it is
considered very likely that the Court of Appeals will rule in
favor of allowing Ms. Hail to compete for promotion.
The support for admitting women to the force on an equal
basis with men is far less than that for admitting ethnic minorities.
The image of the strong, male policeman dies hard. The strength
and agility issues have not been resolved. The rank-and-file
229
patrolman fears that he might have to rely on a "weaker and
more fragile woman" in a life and death situation. There is no
hard evidence that women either can or cannot perform as well
physically as men, but lots of strongly held beliefs. There is
also a powerful, culturally derived reluctance among policemen
to work intimately with women. If a man and a woman were
partners, they would be riding together in a patrol car for long
hours, day after day. Sexually compromising situations are
feared. Policemen's wives are unhappy. The issues are deep
and powerful. Their resolutior remains to be seen. The sub-
stantive changes that a fully integrated and representative police
force might bring to police work are fascinating to contemplate.
L2SSONS FROM TURNOVER EXPERIENCE
The other side of recruiting and selecting people is re-
taining them as members in good standing of the police force.
In our brief discussion of turnover experience in Chapter 2, we
found that of the officers appointed between August, 1962, and
August, 1971, 25, or 4. 25 percent, were fired and 144, or
24. 48 percent, resigned (see Table 2-6). Of the Oakland
residents, 20. 2 percent resigned; of the other California
residents, 19. 3 percent resigned; of the out-of-state residents,
43. 6 percent resigned (see Table 2-7). The residence require-
rent and shift in emphasis to local recruiting should certainly
decrease the number of resignations,
Broken down by educational level, 16. 51 percent of the high
school graduates, 25. 56 percent of those with less than two years
of college, 19. 38 percent of those with more than two years of
college. and 45.83 percent of college graduates resigned (see
Table 2-8). This is consistent with our finding in Chapter 4 that
police salaries are U2 percent of the mean income of high school
graduates and only 79 percent of the mean income of men with
four years of college (see Table 4 -9). Yet Chapter 5 showed
that applicants with more education were more likely to be
selected, at least with the old written tests.
The implication here is that if the Oakland Police Depart-
ment wishes to attract and retain college graduates, perhaps
some system of differential rewards for higher education should
be offered. Similarly, if the department wants to encourage re-
cruits to go back to college and continue their education, perhaps
more than incentives during schooling is needed. Increased re-
wards after the schooling is completed may be necessary as
well.
It is true that college graduates have a better chance of
being promoted than others. Promotional examinations are
231
written tests similar to those discussed in Chapter 5. People
with higher levels of education should have a better chance of
passing them than others. Promotional opportunities, however,
are limited.
The more difficult question is whether educational differ-
entials should be built into the salary structure at a given rank.
Should patrolmen with more education get paid more? Would this
help to retain those who might be apt to resign? Even if so, would
there be overriding disadvantages to a system of differential pay
as well?
We seek some evidence on these questions from a detailed
analysis of the 163 departmental resignations between July, 1966,
and June, 1971. The individuals who resigned are grouped by
years of education in Table 6-5.
Table 6-6 shows the category of position they took after
resigning, by years of education. Of the total group, 47 individ-
uals, or 28.7 percent, went to another state or local police
agency: another 11 individuals, or 6.8 percent, went to the
Federal Bureau of Investigation or the Federal Bureau of Nar-
cotics. Over one-third of the total individuals resigning, there-
fore, remained in the law enforcement field. Of the 34 officers
with B. A. degrees, 18, or 52.8 percent, remained in law en-
232
Table 6-s. Educational level of individuals who resigned fromthe Oakland Police Department, July, 1966, to June, 1971
Education Number Percent
Less than high school* 5 3.1High school graduate 36 22.1Less than 2 years junior college 25 15.3More than .2 years junior college 11 6.7A. A. degree 3 1.8Less than 2 years college 18 11.02 to 4 years college 31 19.0B.A. or B.S. degree 34 20. 9
Total
Source: Oakland Civil Service Office
163 100.0
Note: *Before the end of World War II, some people wereadmitted with less than a high school education.
forcement; of 18 officers with less than two years of college,
eight, or 44.6 percent, remained in law enforcement; of 11
officers with more than two years of junior college, 6, or 54.5
percent, remained in law enforcement. On the contrary, of the
36 high school graduates, eight, or 22.3 percent, remained in
law enforcement; and of the 25 with less than two years of
junior college, four, or 16 percent, remained in law enforcement.
Over half the college graduates, therefore, went to more
attractive opportunities within the field of law enforcement.
None had been promoted in Oakland. Seven of these went to the
Tab
le 6
-6.
Typ
e of
job
take
n by
indi
vidu
als
who
res
igne
d fr
om th
e O
akla
nd P
olic
e D
epar
tmen
t,Ju
ly, 1
966,
to J
une,
197
1, b
y ed
ucat
iona
l lev
el
Typ
e of
Job
or R
easo
n fo
rT
otal
Lee
s th
anH
igh
Scho
ol
Hig
hSc
hool
Gra
duat
e
Les
s th
an2
Yea
rsJu
nior
Col
lege
Mor
e th
an2
Yea
rsJu
nior
Col
lege
A. A
.D
egre
e
Les
s th
an2
Yea
rsC
olle
ge
Mor
e th
an2
Yea
rsC
olle
ge
B. A
.or
B. S
.D
egre
eR
esig
ning
No.
%N
o./0
No.
%N
o.N
o.%
No.
%N
o.%
No.
%N
o.%
Tot
al16
310
0.0
510
0.0
3610
0.0
2510
0.0
1110
0.0
310
0.0
1810
0.0
3110
0.0
3410
0.0
Bay
Are
alo
cal p
. d.
2213
.5
240
.0
411
.1
14.
05
45. 4
527
.82
6. 5
38.
8O
ther
Cal
if.
loca
l p. d
.9
5.5
120
.0
25.
61
4.0
19.
13
9.6
12.
9C
alif
orni
a H
igh-
way
Pat
rol
31.
81
4.0
15.
61
3.2
Out
-of-
stat
elo
cal p
. d.
21.
22
11. 2
Out
-of-
stat
est
ate
p. d
.10
6.1
25.
61
4.0
13.
26
17. 6
Cam
pus
y. d
.1
.6
12.
9
Cum
ulat
ive
tota
l, st
ate
and
loca
l pol
ice
agen
cies
4728
.7
360
.0
822
.3
416
.0
654
. 58
44. 6
722
. 511
32. 2
FBI
53.
15
14. 7
Fede
ral B
urea
uof
Nar
cotic
s6
3.7
412
.92
5.9
Tab
le 6
-6, c
ontin
ued
Typ
e of
Job
or R
easo
n fo
rR
esig
ning
Tot
al
Les
s th
anH
igh
Scho
ol
Hig
hSc
hool
Gra
duat
eN
o.%
No.
%N
o.%
Cum
ulat
ive
tota
l, al
l law
enfo
rcem
ent
agen
cies
5835
. 53
60. 0
822
.3
Bay
Are
alo
cal f
ire
d.4
2. 5
25.
6O
ther
Fed
eral
agen
cies
42.
5T
each
ing
21.
2
Cum
ulat
ive
tota
l, al
l pub
-lic
em
ploy
men
t 68
41.7
360
. 010
27.9
Priv
ate
busi
ness
3521
. 54
11.1
Atte
nd s
choo
l9
5. 5
25.
6M
ilita
ry6
3. 7
Pers
onal
2817
.21
20. 0
1233
. 3A
sked
tore
sign
127.
45
13.9
Ret
ired
53.
11
20. 0
38.
3
Sour
ce: O
akla
nd C
ix it
Ser
vice
Off
ice
Les
s th
an2
Yea
rsJu
nior
Col
lege
No.
%
416
. 0
416
.0
1040
.01
4. 0
14.
06
24. 0
312
.0
Mor
e th
an2
Yea
rsJu
nior
Col
lege
A. A
.D
egre
e
Les
s th
an2
Yea
rsC
olle
ge
Mor
e th
an2
Yea
rsC
olle
ge
B. A
.or
B.S
.D
egre
eN
o.N
o.%
No.
%N
o.N
o.c-
c,
654
. 58
44.6
1135
.418
52 8
133
. 31
3. 2
133
. 33
9.7
15.
61
2. 9
654
.52
66.6
950
.215
48.3
1955
.7
218
.26
33.3
619
.47
20.6
19.
12
6. 5
38.
83
9.7
25.
91
9. 1
133
. 31
5. 6
516
. 11
2. 9
19.
11
5.6
25.
91
5. 6
235
prestigious Fede ral agencies, the FBI and the Bureau of Nar-
cotics. Of the seven, one had six years on the force, two had
five years experience, three had four years, one had three
years. Of the seven, four were from out-of-state, two from
elsewhere in California, only one from Oakland. This is clear
evidence of the risk of trying to be part of a national, professional
labor market. For people whose primary loyalties are to a
national law enforcement professional career, it would be diffi-
cult for Oakland to compete with the FBI.
This conclusion is reinforced if we look further behind the
statistics. Six more of the 18 college graduates who remained in
law enforcement went to out-of-state police agencies; five of
these, all originally from Florida (two with five years of
service, two with four, one with three) went to the Florida Bureau
of Law EnforceMent. The other, originally from Berkeley, be-
came Director of Law Enforcement for the State of Illinois, after
three years of service in the Oakland Police Department. We
have now accounted for 13, over two-thirds, of the college grad-
uates who resigned from the Oakland Police Department and re -.
mained in law enforcement.
What about the other five? One became an Assistant Chief
of Police with the Northern Arizona University Campus Police
236
Department, after six years of experience in Oakland. One,
originally from Fresno, joined the Bakersfield Police Depart-
ment after five years of Oakland service. Oakland's top patrol-
man salary was $932 a month in 1968; Bakersfield's top salary
was $874 a month. 51 But he was not promoted in Oakland;
perhaps he was promoted in Bakersfielu.
The remaining three joined the Walnut Creek Police De-
partment, two (from out-of-state) after five years of Oakland
service, one (from Oakland) after four years. Walnut Creek is
essentially a quiet Bay Area suburb of 33,850 people, It has a
b2-man police force and requires 60 units of college credit.
Although its 1968 starting salary, $801 a month, was lower than
Oakland's $874 a month, its top patrolman salary was $974,
higher than Oakland's $932. 52 Even at the same salary, Walnut
Creek is a far easier place in which to work than Oakland, with
all the problems of an urban core city. The men may have been
promoted as well.
We have explained, therefore, why 18 college graduates
resigned from the Oakland Police Department but remained in
law enforcement work. They went to prestigious Federal
agencies, to home state agencies, to possible promotions and to
quieter jobs. Could we argue convincingly that if the incentives
237
of more money or of promotion had been available they would
have remained? Not unambiguously, because we cannot separate
out taste factors and pecuniary ones to determine at what price
the apparent distaste for Oakland could have been bought off.
Returning to a home agency or leaving a core city may be quite
powerful emotions. Perhaps those who followed their call would
not have continued to be good Oakland patrolmen and should not
have been encouraged to stay.
What about the 16 college graduates who left the law en-
forcement field after they resigned' Seven went into private
business, about which we have no further information. Three
left to attend school and one to teach. Two left to join the mili-
tary, two were asked to leave, and one left for personal reasons.
Again, we have no evidence about the relative power of tastes
and money and no evidence that these men should have been en-
couraged to stay for Oakland's benefit.
Of the 45 men with over two years of junior college or
college or an A.A. degree, 17 resigned and rerliained in law
enforcement occupations. Four of these joined the Federal Bur-
eau of Narcotics, only one went out of state, to the Alaska State
Police. The remaining 12 men, all with one to six years of
service, joined California police departments. The departments
238
chosen and their 1968 top monthly patrolman salary are listed
here (the Oakland salary was $932 per month): San Jose Police
Departmen, ($944); Contra Costa Sheriff's Department ($927);
Concord Police Department ($927); Fremont Police Department
($927-2 men); California Highway Patrol ($893); Fresno County
Sheriff's Department ($855); Pleasanton Police Department
($841); South Lake Tahoe Police Department ($841); Modesto
Police Department ($819); and the Pleasant Hills and Livermore
Police Departments (no salaries available). 53
For this group of observations, one can make a stronger
case that greater economic incentives or promotional opportu-
nities in Oakland may have compensated for the disadvantages
of central city service. San Jose, with a 60 units of college
credit requirement, was the only department that actually paid
more than Oakland, but Contra Costa, Concord and Fremont
came within $5 a month and are far easier and more pleasant to
police. We have no information, moreover, about whether pro-
motions were involved in any of these changes, particularly
likely in the smaller, lower paying departments mentioned. The
potential for pecuniary and promotional incentives to have made
a difference seems better for this group than for the college
graduates. The men remained closer to home and seemed to be
239
motivated primarily by a desire to leave Oakland, rather than
to pursue a more prestigious opportunity. How much of a
differential might have been necessary to overcome the taste
for suburban or rural departments is, of course, unknown.
Of the 18 men with less than two years of college who
resigned, eight remained in the law enforcement field, a per-
centage of 44. 6. Two went out of state to small town depart-
ments in Minne3ota and Washington. Two joined the San Fran-
cisco Police Department, then paying a top salary of $983, $50
a month higher than Oakland. The others went to Concord,
Fremont, Hayward ($862), and the California Highway Patro1.54
Again, for this group, it appears that improved economic
opportunities within Oakland might have made a difference.
Only 12 of the 61 men with less than two years of junior
college or a high school degree remained in law enforcement
occupations, a lower percentage than in the more educated
groups. Three went out of state, nine remained in California
departments: Fremont; San Leandro ($927): the California
Highway Patrol; three to the Marin County Sheriff's Department
($870) with substantial promotions; South Lake Tahoe; and the
Ventura and Sierra County Sheriff's Departments (no salary55information available). There were a greater number of
240
resignations for personal reasons and men asked to resign in
this group than in the groups with more education. Economic
opportunities may well have convinced those who joined other
local departments to remain.
Does following up the futures of the Oakland resignees
yield any revealing patterns') For the more than one-third who
remained in the law enforcement field, one must ask why Oakland
lost men to competing agencies and whether or not changes in the
economic incentives or promotional opportunities open to Oak-
land patrolmen might have retained them. There seem to be
several overlapping types of labor markets here. One is a
national, professional labor market of college educated law en-
forcement men, mobile, seeking the best available professional
opportunities, like college professors or young business school
graduates. Because of Oakland's elaborate national recruiting
campaigns at colleges in the 1960's and because of its excellent
reputation as a professional department, the city could compete
in this market. It offered initial on-the-job training and patrol.
experience. However, for this group of professionals, if more
prestigious opportunities appeared, on they moved. To retain
them, the city would have had to provide equivalent upward
mobility channels, either through establishing more higher
241
level positions which could be attained more quickly by college
graduates, or through significant pay differentials for education.
Educational pay differentials, however, would have serious
costs, not merely budgetary. Comraderie, team loyalty, equality
play a vital role in the functioning of police departments. Accen-
tuating divisions among men by differential incentives based on
education may undermine the principles upon which the organiza-
tion functions. Local school systems and major universities have
faced.a similar problem. Should teachers of subjects with more
attractive alternative opportunities get paid more than others?
Should teachers of more difficult or more prestigious subjects
get paid more than others? The educational field seems to
accept higher pay for graduate school credits or Master's
degrees, but resists further differentiation on economic gro.unds.
A fine economic case can be made for differential incentives for
teachers in shortages fields to obtain them without raising the
pay of teachers who can be attracted at lower salaries. 56 The
trend in school systems, on the contrary, is towards equalizing
salary schedules and employee organizations prod further in
this direction. Differentiation among teachers on economic
grounds is repugnant to the very values upon which school sys-
tems are based, in a way very similar to that in which differ-
242
entiation among policemen on educational or economic grounds
would be viewed by police organizations. The institutional re-
quirements of harmonious and equal working relationships among
colleagues may far outweigh economic rationality. It is not even
clear that economic incentives would work or that Oakland could
hope to compete with the FBI.
If an unambiguous case cannot be made for offering law
enforcement professionals incentives to stay, should they be
hired at all? Here again, a clear judgment is difficult. Per-
haps the intangible benefits to the department, in high level
performance and professional tone, outweigh the investment lost
when the men leave. With the abandonment of national'recruiting
and the establishment of residence preference, fewer men will
be hired from a national market and these questions are likely
to become moot. It will be interesting to observe the turnover
experience with the local college graduates that are hired.
There are two local labor markets, majority group and
minority group, each with a range of educational experience.
Caucasians who resign and remain in law enforcement are most
likely to leave the core city for quieter local positions, often
with better promotional opportunities. The relevant economic
incentive to retain this group would be larger pay differentials
243
over suburban or rural cities to compensate for the battle scars
and anxiety of big city duty, as well as more promotional
opportunities for experienced patrolmen. Increased vacation
opportunities or sabbatical leaves to provide the refreshment of
distance from tension are often suggested as well. These latter
proposals may have independent merit for all policemen, but not
for the city treasury. The men who leave for more peaceful
cl2partments, however, may be just the men who should be en-
couraged to leave, because they have lost their desire to cope
with Cakland, its problems, and the nonpecuniary stresses of the
job. This cannot always be predicted in advance, and turnover
costs here may be well worthwhile.
There has been too little experience with the local minority
community labor market to be able to evaluate its turnover. Mi-
nority group members woult be less likely to be lured by the
attractions of small enclaves and more likely to be committed to
core city problems. Incr.. g the minority composition of the
force may well have signifi.:.ant side benefits in reducing overall
turnover, particularly if the general hostility between the police
and minority communities can be reduced.
What about the almost two-thirds of the resignees who
leave the law enforcement field? The largest group, 35 indi-
244
viduals, 21. 5 percent of all resignations, left to enter private
business; the next largest group, 28 individuals, 17.2 percent
of resignations, left for personal reasons. We have rio more
det- fled information about the jobs they took. We can only infer
that a more attractive combination of pecuniary and nonpecuniary
incentives emerged for these people outside of law enforcement;
we can say nothing about the relative pushes and pulls of money
and tastes. It is interesting to note that the less educated groups
were more likely to leave for personal reasons and less likely
to remain in law enforcement work than others, suggesting per-
haps that college exposure enables people to make more informed
and better career choices. Included in the group who left for
personal reasons are those unhappy with the demands of the life-
style of a policeman--night work, marital and family pressures,
and general tensions. Again we can argue that such turnover
should not be discouraged; indeed the unhappy or unsuited should
be encouraged to leave.
Our speculations about resignations have led us to con-
clude that educational pay differentials would not be a desirable
means of reducing turnover, Some of this turnover will be
taken care of by recruiting Oakland residents and minorities.
The possible effectiveness of differentials in retaining highly
245
educated men would be swamped by the morale effects on the
police organization. Given the nature of the police job, some
turnover has its value, particularly in so far as people with
distaste for the city or the work leave. Perhaps we should be
more concerned about people who remain on the force, despite
dissatisfaction, because of a lack of attractive alternative
opportunities or motivation. In any case, educational differen-
tials are likely to create more problems than they solve.
Substantive restructuring of the job, however, to make it
more flexible and to better match duties with individual skills
and interests, may well be called for. More promotional
opportunities may well be provided in the context of a general
restructuring of the job. Perhaps the changing composition of
the force, adding women as well as minorities, will provoke such
changes. To do justice to the rich possibilities here, an addi-
tional study would be required.
The monopsony hypothesis is contradicted once more by
the evidence gleaned from our study of resignations. Indeed,
there is competition among police departments. Big city status
seems to yield not monopsony power to offer low wages, but the
need for higher wages than surrounding departments to compen-
sate for the disutilities involved in big city duty. Roger W.
246
Schmenner's empirical work, discussed in Chapter 3, gave him
the suspicion that this might be true. 57 Our discoveries offer
some confirmation. Police departments do not seem to wield
monopsony power.
247
CHAPTFR
FOOTNOTES
1. U. S. Department of Commerce, Social and EconomicStatistics Administration, Bureau of the Census, 1970 Censusof Population, General Social and Economic Characteristics,California.
Z. Oakland Civil Service Office.
3. United States District Court, Northern District of California,Clarence Penn et al. vs. William Stumpf, et al. , CivilAction No. C-69-239 OJC.
4. City of Oakland, "Inter-office Letter from City Attorney Ed-ward A. Goggin and Deputy City Attorney William C. Sharpto James M. Newman, subject Penn v. Stumpf," April 10,1972.
5. Clarence Penn vs. William Stumpf (1970) 308 Fed. Supp.1238.
b. Chief Justice Warren Burger, Willie S. Griggs v. DukePower Company, 28 L. ed 2d 158, paragraph 12.
7 Equal Employment Opportunity Commission, Guidelines onErn_ployee Selection Procedures, Federal Register, Volume35, Number 149, August 1, 1970, Washington, D. C. , pp.12333-1 to 12333-3.
8. , pp. 12333-1.
9. Ibid.
1'). .bid.
11. Cr.,i1 Rihts Act of 19(4 (42 U. S. C. , 2000e, et. seq. ),Section 703 (a) (2).
12. Equal Employment Opportunity Commission, Guidelines onEmployee Selection Procedure, 22. cit. , p. 12333-1.
248
13. American Psychological Association, "Standards for Educa-tional and Psychological Tests and Manuals," (Washington,D.C. : American Psychological Association).
14. Equal Employment Opportunity Commission, Guidelines cryEmployee Selection Procedures, 22. cit., pp. 12333-2
15. For an excellent summary discussion of the case law, seeRichard F. Biddle, Discrimination: What Does It Mean?(Chicago: International Personnel Management Association),
16 16 pp. Also see State of California, Fair EmploymentPractices Commission, Guidelines on Employee SelectionProcedures, October, 1972: and Betty R. Anderson andMartha P. Rogers, Personnel Testing and Equal EmploymentOpportunity (Washington, D. C.: U. S. Government PrintingCffice, December, 1970).
16. Richard E. Biddle, Discrimination: What Does It Mean? ,2. cit. , p. 7 referring to Moody v. Albermarle, Fourth
Circuit Court of Appeals.
17. William K. Hunter and Thomas J. Culligan, "An Analysisof Factors to Predict Work Performance of Oakland PoliceOfficers," James M. Newman, Assessment of PoliceRecruitment, Selection and Training: An InterdisciplinaryApproach, Final Report: Recruitment and Selection (Berkeley,California: Education/Research, Inc. , February, 1972).
18. The middle individuals were eliminated because their in-clusion would cloud the process of discriminating betweentop and bottom performers and because the differences inthe raw scores on which the rankings are based was notnearly as distinctive for the middle 60 as for the top andbottom 30. Two more were eliminated due to lack of testscores and background information. See William K. Hunterand Thomas J. Culligan, ibid. , pp. vi-3.
19. Ibid. , Table 1, pp. vi-10.
20. Ibid.
21. William K. Muir, "Personal Characteristics that Contributeto Good Police Work, James M. Newman, Assessment of
249
Police Recruitment, Selection and Training: An Inter-disciplinary Approach, Final Report: Recruitment andSelection, 22. cit.
22. Ibid. , pp. vii-33.
23. Ibid. , pp. vii-37.
24. Ibid. , pp. vii-38.
25. Ibid. , pp. vii-42.
26. Ibid. , pp. vii-43.
27. Dr. Floyd 0. Due, "History and Personality CharacteristicsFavoring Success as a Police Officer," unpublished paper,1972.
28. Talk with James M. Newman, Personnel Director, City ofOakland, July 17, 1974.
29. State of California, Fair Employment Practices Commission,Affirmative Action Survey Thy California FEPC, May, 1974,58 pp. , unpublished paper.
30. Ibid.
31. Ibid.
32. United S.trates District Court, Northern District of California,Clarence Penn et al. vs. William Stumpf et al. , Civil ActionNo. C-69-239 OJC, Consent Decree December 20, 1973,p. 5.
33. Oakland Civil Service Office.
34. State of California, Fair Employment Practices Commission,Affirmative Action Survey by California FEPC, 22.. cit.
35. Ibid.
36. Oakland Civil Service Rule 4:03, July 20, 1971.
250
37. See Silas Ector v. City of Torrance, 2 Civil No. 39769,California Court of Appeals.
38. Oakland City Council Resolution No. 51836, October 26, 1971.
39. United States District Court, Northern District of California,Clarence Penn et al. vs. William Stumpf, et al. , Civil ActionNo. C-69-239 OJC, Consent Decree, December 20, 1973,P. 3.
40. Ibid.
41. Ibid.
42. Oakland Civil Service Office, "Interim Progress Report onMinority Recruiting Program for the Bay Area," unpublishedpaper.
43. Ibid.
44. James M. Newman, July 17, 1974.
45. Oakland City Council Resolution No. 41933, OCCB, andOrdinance No. 8850 C. M. S. , February 1, 1974.
46. State of California, Court of Appeals, First AppellateDistrict, Division Three, Veragene Hardy vs. William F.Stumpf, 1 Civ. 32689, March 15, 1974.
47. Ibid.
48. Federal District Court, Joan Hail vs. James S. White,g st 2, 1973.
Ibid.
50. J. Stanley Pottinger, Assistant Attorney General, "Brief
Appeals, Ninth Circuit, April 5, 1974, 29 pp.
51. State of California, Commission on Peace OfficersStandards and Training, A Career in Law Enforcement,
Hail vs. James S. White, " No. 74-1038, U. S. Court offor the United States as Amicus Curiae, in case of Joan
October, 1969.
251
52. Ibid.
53. Ibid.
54. Ibid.
55. Ibid.
56. Joseph A. Kershaw and Roland N. McKean, TeacherShortages and Salaries (New York: McGraw Hill, 1962).
57. Roger W. Schmenner, "The Determination of MunicipalEmployee Wages," Review of Economics and Statistics,Vol. LV, No. 1, February, 1973, pp. 83-91.
252
CHAPTER 7
SUMMARY AND CONCLUSIONS
Our search for an explanation of and a solution to the
alleged shortages of police manpower in the 1960's has led us
to three distinct theoretical models.
First, the simplest economic solution to the problem is
to raise relative wages of policemen sufficiently to eliminate
the shortage. The slightly richer Arrow and Capron concep-
tion of dynamic shortages adds that information gaps must be
filled and institutional decisions made, so the wage adjustments
might take some time. If demand continues to increase faster
than supply, moreover, a shortage could be sustained indefi-
nitely; the adjustment process would be an ongoing one.
Secondly, it is argued, most persuasively by Eugene J.
Devine, that local governments are monopsonistic employers
of certain categories of labor, including policemen. Therefore,
shortages in these categories arise because a monopsonist will
experience job vacancies in equilibrium, at a monopsony wage,
and will neither raise wages nor recruitment costs sufficiently
to eliminate them. The shortage, then, is a function of market
power and an equilibrium result.
253
Thirdly, in the most rigorous treatment, Ronald Ehren-
berg stresses the complex interaction between wages, hiring
standards, authorized employment levels, and vacancy rates.
He combines them in a utility maximizing model and emerges
with a set of formal relationships among them, which are
spelled out clearly but are too complex to yield simple conclu-
sions. He explores, as well, the problem of selecting employees
from heterogeneous applicants, given authorized employment and
wage levels, by choosing a minimum acceptable quality level,
defined perhaps by a vector of hiring standards. He specifies
theoretically how to choose this minimum acceptable quality
level in order to maximize the expected value of the net gain to
society. He presents in detail the effect of changes in wages,
number of applicants, authorized employment level, or value
of the public service on the minimum acceptable quality level
and on the expected vacancy rate.
How useful are these models in understanding the exper-
ience of the City of Oakland in the 1960's and 1970's? Particu-
larly high vacancy rates of 6. 3 percent in 1966 and 6.1 percent
in 1968 existed (see Table 4-U). Complaints of shortages of
qualified applicants were loud and persistent, in spite of well
developed national recruiting campaigns and starting salaries
254
equal to 113 percent in 1966 and 117 percent in 1968 of the average
earnings of San Francisco-Oakland area production workers in
manufacturing (see Table 4-7). Did the apparent shortages re-
flect a disequilibrium situation which would be resolved by
relative wage increases? Did they reflect a monopsonistic
equilibrium? Did they emerge from the definition of quality
inherent in the selection process? Can we choose one explana-
tion, or a combination of explanations, or none?
Our evidence is consistent with a dynamic shortage ex-
planation, in which the lags in market adjustment were ultimate-
ly corrected. In 1968, patrolmen's starting salaries were in-
creased twice, 8. 7 percent in July and 6. 5 percent in August.
The starting salary relative to production workers' wages rose
from 117 percent in July, 1968, to 125 percent in July, 1969, to
135 percent in July, 1970, and back to 125 percent thereafter.
Simultaneously, general economic conditions worsened, poli-
tical turmoil lessened, departmental image improved, and
attempts to attract more minorities to the force were launched.
The vacancy rate declined from 6.1 percent of authorized posi-
tions in July, 1968, to zero in June, 1969, and remains very
small. The complaints have shifted from general shortages of
qualified applicants to shortages only of qualified minority
applicants.
255
The implications of conventional economic postulates are
supported as well by the removal of 20 authorized patrolman
positions in July, 1974, a classic employment response to con-
tinued salary increases brought about by the pressures of infla-
tion. The demand for public employees may well be responsive
to wage increases.
We reject the monopsony explanation that the shortages of
the 1960's were equilibrium positions at a monopsony wage.
There is no eviderce of the exercise of monopsony power. Po-
lice departments compete for ambiguously defined qualities like
trainability and good character with a wide range of private
employers. Policemen do not possess specific, unique skills
or licenses which can only be sold to police departments. More-
over, core city police departments compete with surrounding
suburbc,n and rural departments. Their status does not yield
monopsony power to offer lower wages, but demands compen-
sating differentials for the greater difficulties and tensions in-
volved.
We believe, however, in greater complexity than the
simplest economic explanation allows. We support Ehrenberg's
emphasis on the interactions among wages, hiring standards,
authorized employment levels and vacancy rates. We would
256
add recruitment and training as well. Ehrenberg stresses the
theoretical importance of the minimum acceptable quality level
chosen, but dismisses its substance as "not directly observable."
We have concentrated on observing and defining the vector of
hiring standards of the Oakland POlice Department, and the
changes therein, between 196r. and 1974. We have found that the
particular recruitment sources, hiring standards and selection
processes used determine directly the available labor supply
and the substantive composition of the police force.
The policing job does not lend itself to clear economic
analysis by specifying a production function, determining derived
demand functions, and stating the qualities required of the labor
supply. It is not only a public service, with all the attendant
difficulties. of defining prices and outputs, but it is a politically
sensitive public service with inherently contradictory elements.
The ambiguities of the job, and, therefore, of the qualities
necessary for its successful performance, are discussed in
Chapter 2.
Nevertheless, prevailing ideas about the minimum accept-
able characteristics of policemen were translated into detailed
requirements about age, height, weight, vision, and education,
and screened further by Mental Adaptability, Civil Service,
257
Physical Agility, and Oral selection examinations, all of which
seriously restricted the available labor supply. The Oakland
Police Department's Personnel Division estimated that less
than half of one percent of Oakland's total population in 1969 was
eligible to apply and approximately one-tenth of one percent
would pass successfully through the selection process. If the
male population 21 to 32 years old is used as a base, still only
slightly over 2 percent of that age group would succeed in the
tests. Clearly the qualifications and selection devices chosen
were partly responsible for the shortage complaints.
What characteristics were sought by the Mental Adapta-
bility tests, the Civil Service test, the Physical Agility test, and
the Oral examination? Using data from 2, 366 employment appli-
cations submitted between 1965 and June, 1971, and multiple
regression techniques, we investigated the determinants of
success in the selection process. We find, for example, that
while high school graduation or its equivalent is the only formal
educational requirement, the probability of success in the Men-
tal Adaptability is increased by large magnitudes as years of
higher education increase from some years of junior college
through three to four years of senior college or a college de-
gree. Senior college education still has a significant effect,
moreover, on the ability to pass the Civil Service test.
258
Local residence appears on balance to have a negative
effect on the probability of selection in these years of national
recruiting. Out-of-state candidates, particularly those found
on college campuses, would be functioning in a national, pro-
fessional labor market for law enforcement personnel, seeking
out the best career opportunities. They would be likely to have
invested in more search and more self-selection before applying,
and, therefore, to be more motivated or interested than the
average local resident. Turnover figures show, however, that
out-of-state residents would also be more likely to resign than
California residents (see Table 2-7) to take better professional
opportunities elsewhere.
Minority status, whether measured directly or indirectly
by the proxy of high school attendance in the South, had a large
and significant negative effect on the probability of passing the.
written tests. In 1966, only 3 percent of the Oakland police
force was minority: by the end of 1970, the percent minority
had increased to 6. 5 (see Table 6-2). The small minority
representation seemed to be a direct result of recruitment and
selection policies which reinforced the suspicious attitude of
the minority community towards the police, an alien organiza-
tion.
259
The pressures of the 1970's have produced policy changes
which have significantly increased minority representation on
the force to 16 percent of the whole force and 61. 0 percent of
the new hires in 1973. Moreover, the Oakland experience has
demonstrated that minority representation can be improved
without sacrificing standards of quality. Widely publicized
changes in recruitment emphasis, rendered concrete and
credible by a residence requirement, an Affirmative Action
program, mass media advertising, and demonstrable results,
have clearly changed the composition and the public image of
the department. The Mental Adaptability and Civil Service
general knowledge tests have been replaced by a more content
valid written test, primarily of reading comprehension. Mi-
nority group members are still substantially more likely to
fail the written test than others (see Table 6-4). It wo.Ild be
interesting, in future research, to compare the determinants
of success on the new written test with the determinants of
success on the Mental Adaptability and Civil Service tests be-
tween 1965 and 1971. Yet, in spite of the continued use of a
written test, Oakland, through concentrated efforts on other
fronts, seems well on its way to eliminating the adverse effect
of its selection process on the minority composition of the
260
police force. Even in 1971, those minority members who passed
the written had a higher probability of passing the Oral than
other candidates: this continues to be true. No other device in
the selection process has an adverse effect on minority candi-
dates, and today's t ggressive minority recruiting has clearly
brought positive results.
As for other characterisLics sought by the selection pro-
cess, specific investment by applicants in police science courses
is unimportant, but previous experience as a policeman does
have a significant, positive effect on the probability of success
at. all stages of the process. The role of previous salary as an
indicator to the department of an applicant's potential perfor -
mance is clouded by the difficulties of filtering out partmtime
from full-time work. The highest previous salary category,
$700 a month or more, does have a consistent, significant
positive effect on the probability of a man being accepted. Other
salary groups were inconclusive. If, however, we minimize the
distorting effects of part-time work by looking separately at the
group with no college experience, higher salaries do indicate
progressively greater probabilities of being selected, as econ-
omic theory would predict. The number of previous jobs enum-
erated on the application is unimportant. Age has a.small,
261
negative effect on passing the Mental Adaptability, resulting
from the well-known attrition of test-taking ability with age,
holding education constant. Age, however, has a positive effect
on passing the Cral, reflecting a concern with the candidate's
maturity.
The legal insistence in the 1970's that if a selection device
has an adverse effect on the employment of groups protected by
the Civil Rights Act of 1964, it must be demonstrably job-
related, and the empirical difficulties inherent in proving job-
relatedness, have led women to challenge other standards
hallowed by tradition. The distinction between patrolman and
policewoman has been removed by Oakland through the estab-
lishment of the combined category, police officer. Height and
weight stattlards which would exclude the majority of women
from the police officer job have been abolished by court order.
Another substantial change in the composition of the Oakland
Police Department seems imminent. All hiring has currently
been suspended while a job-related Physical Agility exam which
will adequately screen the strength and physical abilities of
women is developed. How many women will succeed in being
selected and how well women will be able to perform as police
officers remains to be seen. Far reaching cultural adaptations
262
may well be required of the police organization. Perhaps they
will lead to a long overdue substantive restructuring of the job
based on a more careful appraisal of matching particular police
functions and individual skills.
Thoughtful analysis and restructuring of the job may well
yield a conception of a professional police officer based not on
qualities like college graduation, which have only an ambiguous
relationship to job requirements and performance, but on a
clear understanding of the dimensions and demands of the job.
We rejected differential economic incentives to retain more
educated police officers because, in the existing structure, its
threats to morale, equality, comraderies, and organizational
effectiveness seemed to far outweigh any possible gain. Far
better use could be made of economic incentives in the context
of a clear understanding of the performance requirements of
the job.
We have concluded, therefore, that economic incentives,
recruitment and selection must be viewed together. More
attempts must be made to understand their interaction, both in
theory and in practice. Techniques must be developed to
observe, analyze and understand vectors of hiring standards
and their economic functions. In the pol ice case, the vector of
263
hiring standards used defines not only the potential labor
supply, but the nature of the job itself. It will be fascinating
to watch the substantive effects of changes in the vector of
hiring standards on the police job of the future.
BIBLIOGRAPHY
BOOKS
264
Becker, Gary S., Human Capital. New York: Collo-n*4a thaiversityPress, 1964.
Berkeley, George E., The Democratic Policeman. B.osOon:Beacon Press, 1961.
Blank, David H. , and Stigler, George, The Demand and Supply ofScientific Personnel. New York: NBER, 1957.
Blaug, Mark, editor, Economics of Education, 1. Baltimore:Penguin, 1968.
Boudua, David J. , The Police. New York: Wiley, 1967.
Bowen, William G. , and Finegan, T. Aldrich, The Economics ofLabor Force Participation. Princeton, N. J.: PrincetonUniversity Press, 1969.
Chevigny, Paul, Police Power. New York: Vintage Books, 1969.
Devine, Eugene J., Analysis of Manpower Shortajes in LocalGovernment. New York: Praeger, 1969.
Durkheim, Emile, Professional Ethics and Civic Morals. Glencoe:The Free Press, 1958.
Ehrenberg, Ronald G., The Demand for State and Local Govern-ment Employees. Lexington, Mass. : D. C. Heath, 1972.
Franke, Walter, and Sobel, Irvin, The Shortage of Skilled andTechnical Workers. Lexington, Mass. : D. C. Heath, 1970.
Friedman, Milton, Price Theory: A Provisional Text. Chicago:Aldine, 1962,
International City Managers' Association, Municipal Yearbook,1965-1973. Chicago: International City Managers'Association, 1970.
265
Johnston, J. Econometric Methods, 2nd edition. New York:McGraw Hill, 1972.
Juris, Hervey A., and Feuille, Peter, Police Unionism. Lexing-ton, Mass. : D. C. Heath, 1973.
Kershaw, Joseph A. , and McKean, Roland N., Teachers Short-ages and Salaries. New York: McGraw Hill, 1962.
Lecht, Leonard A. , Manpower Needs for National Goals in the1970's. New York: Frederick A. Praeger, 1969.
Niederhoffer, Arthur, Behind the Shield. Garden City, N. Y. :Doubleday, 1967.
Phelps, Edmund S. , editor, Microeconomic Foundations ofEmployment and Inflation Theory. New York: W. W.Norton, 1970.
Radano, Gene, Walking the Beat. Toronto: Collier Books, 1969.
Rees, Albert, and Schultz, George P. , Workers and Wages in anUrban Labor Market. Chicago: University of ChicagoPress, 1970.
Reiss, Albert J., Jr., The Police and the Public. New Haven:Yale University Press, 1971.
Rheinstein, Max. , Max Weber on Law in Economy and Society.Cambridge, Mass. : Harvard University Press, 1954.
Rubinstein, Jonathan, City Police. New York: Farrar, Strausand Giroux, 1973.
Skolnick, Jorome, Justice without Trial. New York: Wiley, 1967.
Vollmer, Howard, and Mills, Donald L. , Professionalization.Englewood Cliffs, N. J. : Prentice Hall, 1966.
Wellington, Harry H. , and Winter, Ralph K. , Jr. , The Unionsand the Cities. Washington, D. C. : Brookings Institution,1971.
266
West ley, Will iam A. , Violence and the Police. Cambridge, Mass. :MIT Press, 1970.
Whittemore, L. H. , Cop' New York: Holt, Rinehart and Winston,1969.
Wilson, James Q., Varieties of Police Behavior. Cambridge,Mass. : Harvard University Press, 1968.
ARTICLES
Alchian, Armen A. , "Information Costs, Pricing and ResourceUnemployment," in Phelps, Edmund S., MicroeconomicFoundations of Employment and Inflation Theory. NewYork: W. W. Norton, 1970, pp. 27-52.
Altman, Stuart H. "Earnings, Unemployment, and the Supply ofEnlisted Volunteers," Journal of Human Resources IV(Winter, 1969), pp. 38-59.
, ''The Structure of Nursing Education and ItsImpact on Supply," in Klarman, Herbert E., editor,Empirical Studies in Health Economics. Baltimore: TheJohns Hopkins Press, 1970, pp. 335-352.
Archibald, George, "The Factor Gap and the Level of Wages,"Economic Record XXX (November 1954), pp. 187-199.
Arrow, K. J. , and Capron, W. M. , "Dynamic Shortages andPrice Rises: The Engineer Scientist Case," QuarterlyJournal of Economics LXXIII(2) (May, 1959), pp. 292-398.
Ashenfelter, Orley, "The Effects of Unionization on Wages inthe Public Sector: The Case of Fire Fighters," Industrialand Labor Relations Review 24(2) (January,1971), pp. 191-202.
Balch, Robert W. , "The Police Personality: Fact or Fiction,"Journal of Criminal Law, Criminology and Police Science63 (1972), p. 106-112.
Bancroft, Raymond L. , "Municipal Law Enforcement, 1966,"Nations' Cities .4 (February, 1966), pp. 15-17.
267
Barbour, George P. , Jr. , "Measuring Local Government Produc-tivity," International City Managers' Association, TheMunicipal Yearbook 1973, pp. 38-44.
Bronfenbrenner, Martin, "Potential Monopsony in Labor Markets, "Industrial and Labor Relations Review 9(4) (July, 1956), pp.577-588.
Carlsson, Robert J. , and Robinson, James W. , "Criticism andComment: Compensation Decisions in Public Organizations,Industrial Relations 9(1) (October, 1969), pp. 111-113.
, "Toward a Public Employment WageTheory, " Industrial and Labor Relations Review 22(2)(January,1969), pp. 243-248.
Carte, Gene E., "Police Representation and the Dilemma of Re-cruitment, " Issues in Criminology 6(1) (Winter, 1971), pp.85-93.
Chow, Gregory C. , "Tests of Equality between Subsets of Co-efficients in Two Linear Regressions, " Econometrica 28(1960), pp. 591-605.
Colarelli, Nick J. , and Siegel, Saul M. , "A Method of PolicePersonnel Selection," The Journal of Criminal Law,Criminology and Police Science 55 (1964), pp. 287-289.
Craft, James A. , "Toward a Public Employee Wage Theory:Comment, " Industrial and Labor Relations Review 23(1)(October,1969), pp. 89-95.
Ehrenberg, Ronald G. , "Municipal Government Structure,Unionization, and the Wages of Fire Fighters," Industrialand Labor Relations Review 27(1) (October, 1973), pp. 36-48.
Fisher, A. C. , "The Cost of the Draft and the Cost of Endingthe Draft, " American Economic Review LIX (Jung 1969),pp. 239-254.
Fisher, Franklin M., "Tests of Equality between Sets of Co-efficients in Two Linear Regressions, An Expository Note, "Econometrica 38(2) (March, 1970), pp. 361-366.
'I
268
Flynn, J. T. , and Peterson, M. , "The Use of RegressionAnalysis in Police Patrolman Selection," Journal ofCriminal Law, Criminology and Police Science 63(4) (1972),pp. 564-569.
Fogel, Walter, and Lewin, David, "Wages Determination in thePublic Sector," Industrial and Labor Relations Review 27(3)(April, 1974), p. 410-431.
Frankel, Laurie, and Allard, Mary Ann, "Personnel Practice ofMunicipal Police Departments, " ICMA Urban Data Servicedeports 4 (August,1972), pp. 1-17.
Freund, James L., "Market and Ufiaii itifltietice on MunicipalEmployee Wages," Industrial and Labor Relations Review27(3) (Apri1.1974), pp. 391-404.
Garbarino, Joseph W. , "Professional Negotiations in Education,"Industrial Relations 7(a) (February, 1968), pp. 93-101,
Geller, Irving B. , "Higher r.ducation and Policemen," Journal ofCriminal L4W, COrninplogy 4pd Police Science 63(1972), pp. 396-401.
Gerwin, Donald, "Compensation Decisional in Public Organizations, I'Industrial Relattp.np 9(3) (February, 1969), pp. 175 -184,
Gordon, Donald F. , and Hynes, Allan, "On the Theory of PriceDynamics, " in Phelps, Edmund S. , M1,croeconornic Founda-tions of Employment and InflatipnTheory. New York: W.W. Norton, 1970, pp. 369-393.
Greenwood, Ernest, "Attributes of a Profession," in Nosow,Sigmund, and Form, William H., Man, World and Society.New York: Basic Books, 1962, pp. 207-218.
Harman, B. Douglas, "Expenditures for Police and Fire Depart-ments, " International City Managers' Association UrbanData Service Reports ?. (eptember, 1970), pp. 1 25.
Hogan, Robert, "Successful Oakland Police: A PersonologicalProfile, " in Newman, James M., and Rogin, William E. ,Assessment of Police Recruitment, Selection and Training:
269
An Interdisciplinary Approach, Final Report: Recruitmentand Selection. Berkeley, Calif. : Education/Research, Inc.February,197Z, pp. III-1 - III-13.
Hoobler, Raymond L. , and McQueeney, J. A., "A Question ofHeight, " The Police Chief `1., (November,1973), pp. 42-49.
Howe, George F. , "Trends in Salaries of Policemen and Firemen,"International City Managers' Association Urban Data ServiceReports 1(7) (July, 1969), pp. 1-18.
Hunter, William K. , and Culligan, Thomas J. , "An Analysis ofFactors to Predict Work Performance of Oakland PoliceOfficers, " in Newman, James M. , and Rogin, William E. ,Assessment of Police Recruitment, Selection and Training:An Interdisciplinary Approach, Final Report: Recruitmentand Selection, Berkeley, Calif. : Education/Research, inc.,February, 1972, pp. VI-1 - VI-37.
Hurd, Richard W. , "Equilibrium Vacancies in a Labor MarketDominated by Non-profit Firms: The 'Shortage' of Nurses,"Review of Economics and Statistics LV(2) (May,1973), pp.pp. 234-240.
Jensen, 011ie, "Cultural Bias in Selection, " in Golembiewski,Robert T. , and Cohen, Michael, editors, People in PublicService. Itasca, Ill. : F. E. Peacock Publishers, 1970,pp. 287-295.
Kane, Walter C. , and Neuhart, Lynn A. , "Police EducationalIncentive Pay Plans," International City Managers'Association Management Information Service 2(L-5) (May,1970), pp. 1-11.
Lawer, Neil, "Developing New Employment Tests for Minorities,"Public Administration Review XXXI (July-August,1971), pp.459-461.
Levy, Ruth J. , "Investigation of a Method for the Identificationof the High-Risk Police Applicant, Can High Risk PoliceApplicants be Identified?, Institute for Local Self Govern-ment, Berkeley, California, July, 1971, pp. 1-58.
270
"Predicting Police Failures, " Journal ofCriminal Law, Criminology and Police Science 58(2) (1967),pp. 265-277.
Lewin, David, "Wage Parity and the Supply of Police and Firemen,"Industrial Relations 12(1) (February,1973), pp. 77-85.
Manson, Don., and Howe, George, "Recent Trends in PoliceCommunity Relations, " International City Managers' Asso-ciation Urban Data Service Reports 2(3) (March, 1970), pp.1- 19.
Matarazzo, J. D., "Characteristics of Successful Policeman andFireman Applicants," Journal of Applied Psychology 48(1964), pp. 123-133.
McAllister, John A., "A Study of Prediction of Measurement ofPolice Performance," Police 14 (1970), pp. 58-64.
McCall, John, and Wallace, Neil, "A Supply Function of First-term Re-enlistees to the Air Force," Journal of HumanResources IV(3) (Summer, 1969), pp. 293-310.
McDevit, Robert J. , "Situational Tests in Metropolitan PoliceSelections " Journal of Criminal Law, Criminology andPolice Science 57 (1966), pp. 99-106.
McNamara, John H., "Uncertainties in Police Work: The Rele-vance of Police Recruits' Backgrounds and Training, " inBoudua, David J., editor, The Police: Six SociologicalEssays. New York: John Wiley and Sons, 1967, pp. 163-250.
Miller, Christine M. , "A Study of the MMPI as a Selection Devicefor the Oakland Police Department," in Newman, James M. ,and Rogin, William E. , Assessment of Police Recruitment,Selection and Training: An Interdisciplinary Approach,Final Report: Recruitment and Selection. Berkeley, Calif.Education/Research, Inc. , February, 1972, pp. IV-1 - IV-21.
Mills, Robert B. , McDevitt, Robert J. , and Tonkin, Sandra,"Situational Tests in Metropolitan Police Recruit Selection,"Journal of Criminal Law, Criminology and Police Science57 (1966), pp. 99-104.
271
Mort3nsen, Dale T. , "A Theory of Wage and EmploymentDynamics, " in Phelps, Edmund S. , Microeconomic Founda-tions of Employment Theory. New York: W.W. Norton, 1970, pp. 167-211.
"Job Search, the Duration of Unemploy-,
ment, and the Phillips Curve," American Economic Review60 (December,1970), pp. 847-861.
Muir, William K. , "Personal Characteristics that Contribute toGood Police Work, " in Newman, James M. , and Rogin,William E. , Assessment of Police Recruitment, Selectionand Training: An Interdisciplinary Approach, Final Re-port: Recruitment and Selection. Berkeley, Calif. :Education/Research, Inc. , February, 1972, pp. VII-1 VII-124.
Mullineaux, Jewel E. , "An Evaluation of Predictors Used toSelect Patrolmen," Public Personnel Review 16, pp. 84-86.
Newman, James M. , Rogin, William E. , Hunter, William K. ,and Vora, Shanti, "Investigation of a Method for Identifica-tion of the High-Risk Police Applicant: An Analysis andInterpretation Based on Available Data," Can High RiskPolice Applicants Be Identified? , Institute for Local SelfGovernment, Berkeley, California, July, 1971.
0i, Walter, "Labor as a Quasi-Fixed Factor," Journal ofPolitical Economy LXX (December,1962), pp. 538-555.
Sanders; Eric P. , "Oakland Police Sample: A DescriptivePsychographic Study," in Newman, James M. , and Rogin,William E. , Assessment of Police Recruitment Selectionand Training: An Interdisciplinary Approach, Final Re-port: Recruitment and Selection. Berkeley, Calif., :Education/Research, lnc. , February, 1972, pp. V-1 - V-33.
Schmenner, Roger W. , "The Determination of Municipal EmployeeWages," Review of Economics and Statistics LV(1) (February,1973), pp. 83-91.
Spencer, Gilmore, and Nichols, Robert, "A Study of Chicago Po-lice Recruits: Validation of Selection Procedures," ThePolice Chief 38(6) (June,1971), pp. 50-55.
272
Tent, Deborah Ann, and Eisenberg, Terry, The Selection andPromotion of Police Officers: A Selected Review of RecentLiterature, " The Police Chief 39(2) (February, 1972), pp.20 - 29 .
Vaughn, Lamar M., and Tenenbaum, Allen, "Oakland's Recruit-ment and Selection Methods: A Historical Overview," inNewman, James M. , and Rogin, William E. , Assessmentof Police Recruitment, Selection and Training: An Inter-disciplinary Approach, Final Report: Recruitment andSelection. Berkeley, Calif. : Education/Research, Inc.,February, 1972, pp. I-1 - 1-79.
Wentworth, Kenneth L., "Development and Use of Written Tests,"in Donovan, J. J. , editor, Recruitment and Selection inthe Public Service. Chicago: Public Personnel Association,1968, pp. 112-122.
Westley, William A. , "Violence and the Police," AmericanJournal of Sociology LIX (July,1953), pp. 34-41.
Whalen, Charles, and Zannini, Susan, "Trends in Police andFire Salaries," International City Managers' AssociationUrban Data Service Reports 3 (October,1971), pp. 1-16.
Wilensky, Harold L. , "Varieties of Work Experience," in Borow,Henry, Man in a World at Work. Boston: Houghton Mifflin,1964, pp. 125-154.
Wilson, James Q. , "Generational and Ethnic Differences amongCareer Police Officers," American Journal oflociologyLXIX(5) (March, 1964), pp. 522-528.
Yett, Donald E., "Causes and Consequences of Salary Differen-tials in Nursing," Inquiry VII (March,1970), pp. 78-99.
, "The Chronic 'Shortage' of Nurses: APublic Policy Dilemma," in Klarman, Herbert E., editor,Empirical Studies in Health Economics. Baltimore: TheJohns Hopkins Press, 1970, pp. 357-389.
273
PUBLIC DOCUMENTS
Anderson, Betty R. , and Rogers, Martha P. , Personnel Testingand Eaual Employment Opportunity. Washington, D. C. :U. S. Government Printing Office, December, 1970.
Baehr, Melaney E. , Furcon, John E. , and Froemel, Ernest C. ,
Psychological Assessment of Patrolman Qualifications inRelation to Field Performance. Washington, D. C. : U. S.Government Printing Office, November 5, 1968.
City of Oakland, California, City Planning Department, Optionsfor Oakland: A Summary Report on the Oakland 701 Project.December,1969.
Civil Rights Act of 1964 (42 U.S. C. , 2000e, et Eq. ), Section 7C3(a) (2).
Cohen, Malcolm S. , Rea, Samuel A. , Jr. , and Terman, RobertI. , "A Micro Model of Labor Supply, " U. S. Department ofLabor, Bureau of Labor Statistics Staff Paper No. 4.Washington, D. C.: U. S. Government Printing Office,1970.
Danielson, William F. , Police Compensation , a report submittedto the President's Commission on Law Enforcement andAdministration of Justice, 1967.
Equal Employment Opportunity Commission, Guidelines onEmployee Selection Procedures, Federal Register, 35(149)(August 1, 1970). Washington, D. C.
Germann, A. C. , "Recruitment, Selection, Promotion and CivilService, " President's Commission on Law Enforcement andthe Administration of Justice. Washington, D. C. : U. S.Government Printing Office, 1967.
National Commission on Law Observance and Enforcement,"Report on the Police " Washington, D. C.: U. S.Government Printing Office, 1931.
National Institute of Law Enforcement and Criminal Justice, LawEnforcement Assistance Administration, U. S. Department
274
of Justice, "Predicting Patrolman Performance" (asummaiy blurb of the Baehr, et al. study).
"Omnibus Crime Control and Safe Streets Act of 1968," PublicLaw 90-351, 90th Congress, HR 5037, June 19, 1968.
President's Commission on Law Enforcement and Administrationof Justice, Task Force Report: The Police. Washington,D. C.: U. S. Government Printing Office, 1967.
President's Commission on Law Enforcement and Administrationof Justice, The Challenge of Crime in a Free Society.Washington, D. C. : U. S. Government Printing Office,February,1967.
State of California, Commission on Peace Officers Standards andTraining, A Career in Law Enforcement. October,1969.
State of California, Commission on Peace Officers Standards andTraining, California Law Enforcement Recruitment Pro-kram. Law Enforcement Assist. Proj. 232, August 31, 1968.
State of California, Department of Human Resources Development,California Labor Market Bulletin. July. 1971, to July, 1973.
State of California, Department of Industrial Relations, Divisionof Labor Statistics and Research, California Labor Statis-tics Bulletin. July, 196L. to July, 1973..
State of California, Fair Employment Practices Commission,Guidelines on Employee Selection Procedures. October,1972.
U. S. Advisory Commission on Civil Disorders, Report on theNational Advisory Commission on Civil Disorders. NewYork: Bantam, 1968.
U. S. Bureau of Employment Security, Dictionary of OccupationalTitles, 3rd edition. Washington, D. C. : U. S. GovernmentPrinting Office, 1965.
U. S. Bureau of the Census, Census of Population: 1970 DetailedCharacteristics, Final Report PC(1)-D6 California, Section
275
1 and Section 2. Washington, D. C. : U. S. GovernmentPrinting Office, 1972.
U. S. Bureau of the Census, 1970 General Social and EconomicCharacteristics, Final Report PC (1)-C6 California.Washington, D. C. : U. S. Government Printing Office,1972.
U. S. Department of Justice, Federal Bureau of InvestigationNational Academy, The FBI Pledge for Law Enforcemer:tOfficers. Washington, D. C. : Federal Bureau of Investi-gation National Academy, 1959.
OTHER SOURCES
American Psychological Association, "Standards for Educationaland Psychological Tests and Manuals." Washington, D. C. :American Psychological Association.
American Trial Lawyers Association, "Crime and Its Cause inLos Angeles." Lancaster, Calif." Golden West PublishingCompany, 1966.
Aussieker, M. W. , Jr., "Police Collective Bargaining." Chicago,Ill. : Public Personnel Association, 1969.
Bay Counties Peace Officers Association and State of California,Commission on Peace Officers Standards and Training,"Symposium: Professional Police: Fact or Fantasy."April 3, 1968.
Berman, Louise, "An Empirical Study of Economic Determinantsof Public Employee Salary Schedules: City Policemen,Firemen, Typists and Stenographers, " unpublished paper,University of California, Berkeley, 1968.
Biddle, Richard E. , Discrimination: What Does it Mean?Chicago: International Personnel Management Association,
City of Oakland, City Council, Oakland City Council ResolutionNo. 41933, OCCB, and Ordinance No. 8850 C. M. S. ,February 1, 1974.
276
City of Oakland, City Council, Oakland City Council ResolutionNo. 51836, October 26, 1971.
City of Oakland, Civil Service Office, Announcement for Patrol-man Examination, September, 1970.
City of Oakland, Civil Service Office, Files.
City of Oakland, Civil Service Office, "Interim Progress Reporton Minority Recruiting Program for the Bay Area, " un-published paper: 1973.
City of Oakland, Civil Service Office, "Oakland Police DepartmentRecruitment and Patrolman Selection Process, " December,1964, unpublished memorandum.
City of Oakland, Civil Service Office, Oakland Civil Service Rule4:03, July 20, 1971.
City of Oakland, Civil Service Office, Employment Application.
City of Oakland, Civil Service Office, "Inter-Office Letter fromCity Attorney Edward A. Goggin and Deputy City AttorneyWilliam C. Sharp to James M. Newman subject Penn v.
" April 10, 1972.
City of Oakland, Police Department, "Demographic Analysis,"unpublished paper, 1973.
City of Oakland, Police Department, Personnel Division, Memofrom Lieutenant J. McArthur, unpublished, 1973.
City of Oakland, "The City of Oakland's Affirmative ActionProgram," unpublished paper, 1973.
Cohen, Bernard, and Chaiken, Jan M., Police BackgroundCharacteristics and Performance. New York: New YorkCity Rand Institute, August,1972.
Deane, R. T. , "Simulating an Econometric Model of the Marketfor Nurses," unpublished Ph. D. dissertation, Universityof California, Los Angeles, 1971.
277
Devine, Eugene J. , "Manpower Shortages in Local GovernmentEmployment," unpublished Ph. D. dissertation, Universityof California, Los Angeles, 1969.
Due, Floyd 0. , "History and Personality Characteristics FavoringSuccess as a Police Officer," unpublished paper, 1972.
, "Psychiatric and Psychological Screeningof Police Applicants to an Urban Department: ComputerAnalysis of 113 Cases," paper delivered to meetings ofAmerican Psychiatric Association, May 13, 1970.
Encyclopedia of Careers and Vocational Guidance, revised edition,IL Chicago: J. G. Ferguson.
Hogan, Robert, "A Study of Police Effectiveness. " Washington,D. C. : Experimental Publication System, AmericanPsychological Association, June, 1970.
Hoover, J. Edgar, "Should You Go into Law Enforcement? "New York: New York Life Insurance, 1961.
Lewin, David, "Wage Determination in Local Government Em-ployment," unpublished Ph. D. dissertation, University ofCalifornia, Los Angeles, 1971.
Mayall, Donald, Human Resources Development for Oakland:Problems and Policies, SRI Project ID-5638. Menlo Park,Calif. Stanford Research Institute, December, 1968.
Newman, James M. , and Rogin, William E. , Assessment ofPolice Recruitment, Selection and Training: An Inter-disciplinary Approach, Final Report: Recruitment andSelection, CCCJ Grant No. 0156. Berkeley, Calif. :Education/Research, Inc. , February,1972.
Newman, James M. , and Rogin, William E. , Assessment ofPolice Recruitment Selection and Trainin: Interim Resorton Training, CCCJ Grant No. 0156. Berkeley, Calif. :Education/Research, Inc. , March, 1971.
New York Times, Wednesday, August 15, 1973.
New York Times, Sunday, December 3, 1973.
-
278
O'Connor, George W. , "Survey of Selection Methods. " Washing-ton, D. C. : International Association of Chiefs of Police,1962.
Otis, Arthur S. , Otis Self-Administering Tests of Mental Ability,Manual of Directions and Key (Revised) for Intermediateand Higher Examinations. New York: Harcourt, Braceand World, 1922.
Otis, Arthur S. , and Lennon, Roger T. , Otis-Lennon MentalAbility Test, Manual for Administration. New York:Harcourt, Brace and World, 1967.
State of California, Fair Employment Practices Commission,Affirmative Action Survey by California FEPC, unpublishedpaper, May,1974.
Thompson, Frank J , "The Politics of Public Personnel Policyin a Core City," unpublished Ph. D. dissertation, Universityof California. Berkeley, 1973.
Tobin, James, "The Application of Multivariate Probit Analysisto Economic Survey Data, " unpublished Cowles FoundationDiscussion Paper No. 1, July 14, 1955, as revised December1, 1955.
Watts, Harold W. , "An Introduction to the Theory of BinaryVariables (or All About Dummies)," unpublished SystemsFormulation and Methodology Workshop Paper 6404, SocialSystems Research Institute, University of Wisconsin,April 23, 1964.
Yett, Donald, "An Economic Analysis of the Hospital NursingShortage," unpublished Ph. D. dissertation, University ofCalifornia, Berkeley, 1968.
COURT CASES
State of California, Court of Appeals, Silas Ector v. City ofTorrance, 2 Civil No. 39769.
State of California, Court of Appeals, Veragene Hardy vs.William F. Stumpf, 1 Civil No. 32689, March 15, 1974.
279
U S Court of Appeals, Ninth Circuit, Joan Hail vs. James S.White, "Brief for the United States vs. Amicus Curiae, '1No. 74-1038 J Stanley Pottinger, April 5, 1974, casepending
U S District Court for Eastern District of Virginia, RichmondDivision, United States of America v Virginia Electricand Power Compaq et al. , Civ. A. No. 638-70-R, May4, 1971
U S District Court. Massachusetts, Julius Arrington et al v.Massachusetts Bay Transportation Authority, Civ. A. No.69-681, December 22, 1969
U S District Court, Minnesota, Fourth Division, Carter et al.vs Hugh Gallagher et al March 9, 1971
U. S District Court, Northern District of California, ClarencePenn et al. vs William Stumpf et al. (1970) 308, Fed Supp.1238
U. S District Court, Northern District of California, ClarencePenn et al vs William Stumpf et al , Civil Action No.C-69-239 OJC, Consent Decree, December 20, 1973
U. S Supreme Court, Willie S Griggs v Duke Power Company,28 L ed Zd 158, March 8, 1971.